Are 99% of COVID-19 Deaths Among the Unvaccinated?

[Mick West]

I found the above meme on Facebook, and reposted it on Twitter. I should relaly not have been surprised, but I got quite a bit of negative reaction from people saying I was spreading disinformation. But I think it's essentially true, even if the 99.2% figure is a little high although the 99.2% figure is out of date . But let's fact-check with sources. Firstly the first five "Who has been vaccinated" - which might seem uncontroversial, but people were still pushing back at some.

1. Donald Trump and every other living US president.
That's Biden, Trump, Obama, Bush, Clinton and Carter.

Article:

Published March 1, 2021
Updated April 17, 2021
Former President Donald J. Trump and his wife, Melania, quietly received coronavirus vaccinations in January before leaving the White House, an adviser said on Monday.

Source: https://www.nytimes.com/2021/03/01/us/politics/donald-trump-melania-coronavirus-vaccine.html

Article:

Jan 11, 2021
Joe Biden on Monday received the second dose of the Pfizer/BioNTech Covid-19 vaccine on camera, as part of an effort by the President-elect's incoming administration to reassure the country of the safety of the vaccines.

Source: https://abc7news.com/joe-biden-covid-vaccine-coronavirus-covid-19-gets/9561849/

Article:

March 11, 2021
The former First Lady took to Instagram today to share that she and her husband, former president Barack Obama, got vaccinated

Source: https://www.harpersbazaar.com/celebrity/latest/a35810004/michelle-obama-barack-obama-vaccine-selfie/

Article:

March 18, 2021
Former President George W. Bush and former First Lady Laura Bush received their COVID-19 vaccinations at Baylor Scott & White Hillcrest Medical Center in Waco

Source: https://www.kwtx.com/2021/03/18/former-president-first-lady-receive-covid-19-vaccinations-at-local-hospital/

Article:

Feb 24 2021
Now that former President Jimmy Carter and his wife are vaccinated against COVID-19, they have returned to one of their favorite things: church.

Source: https://apnews.com/article/baptist-coronavirus-pandemic-jimmy-carter-rosalynn-carter-5d40e69bfc2a317faad05993f1be55bd

Missing Clinton, but more generally:

Article:

March 11, 2021
All living former Presidents Jimmy Carter, Bill Clinton, George W. Bush, Barack Obama and their respective former first ladies -- with the exception of the Trumps -- are part of a newly released ad campaign urging Americans to get the coronavirus vaccine when it is their turn, a push that is aimed squarely at combating vaccine skepticism.

Source: https://www.cnn.com/2021/03/11/politics/former-presidents-vaccine-psa/index.html

2. All 50 governors, both R and D

Article:

June 22, 2021
Texas Gov. Greg Abbott is vaccinated against COVID-19 — as are all other US state governors.

Source: https://www.businessinsider.com/governors-vaccinated-vaccine-hesitancy-free-beer-2021-5

3. Nearly 100% of congress

This is a little tougher, as up-to-date figure are hard to find.

Article:

According to a CNN survey from May, 100% of congressional Democrats and 92% of Senate Republicans had been vaccinated against Covid-19, while only 44.8% of House Republicans were.

Source: https://www.mercurynews.com/2021/06/11/vaccinated-members-can-ditch-face-masks-on-the-house-floor/

While that's nearly 100% of the Senate, it's more like 80% overall. However many Republicans now are simply not saying what their status is. But "Nearly 100% of congress" might be a bit high.

4. 96% of American Physicians.

Article:

June 11, 2021
The American Medical Association (AMA) today released a new survey (PDF) among practicing physicians that shows more than 96 percent of surveyed U.S. physicians have been fully vaccinated for COVID-19, with no significant difference in vaccination rates across regions.

Source: https://www.ama-assn.org/press-center/press-releases/ama-survey-shows-over-96-doctors-fully-vaccinated-against-covid-19

5. 75-80% of the military (soon 100%)
This might initially seem to over-state the figures, as a few weeks ago they were:

Article:

August 9, 2021
The Navy said that more than 74% of all active duty and reserve sailors have been vaccinated with at least one shot. The Air Force, meanwhile, said that more than 65% of its active duty and 60% reserve forces are at least partially vaccinated, and the number for the Army appears closer to 50%.

Source: https://apnews.com/article/coronavirus-vaccine-us-military-requirement-pentagon-3975940c732352f72e41f6e34a3a2669

However

Article:

Members of the U.S. military will be required to get the COVID-19 vaccine beginning next month

Source: https://apnews.com/article/coronavirus-vaccine-us-military-requirement-pentagon-3975940c732352f72e41f6e34a3a2669

So with the 100% target set for just two weeks from now, the 75% number has probably been reached by now.

Finally the most important number:

99.2% of all people currently dying of Covid-19 (are NOT vaccinated)

Multiple sources give this figure for the unvaccinated deaths in June, usually as a quote from Dr Fauchi on Meet the Press:

Article:

07/05/21 04:42 PM EDT
Anthony Fauci on Sunday said more than 99 percent of the people who died from COVID-19 in June were not vaccinated, calling the loss of life "avoidable and preventable."

"If you look at the number of deaths, about 99.2 percent of them are unvaccinated. About 0.8 percent are vaccinated.

Source: https://thehill.com/policy/healthcare/561585-fauci-more-than-99-of-people-who-died-from-covid-19-in-june-were-not

However the same number came from an AP analysis of government data for May

Article:

only about 150 of the more than 18,000 COVID-19 deaths in May were in fully vaccinated people. That translates to about 0.8%, or five deaths per day on average.

Source: https://apnews.com/article/coronavirus-pandemic-health-941fcf43d9731c76c16e7354f5d5e187

More detailed figures, with some discussion of the challenges, can be found At the Kaiser Family Foundation

Article:

The share of deaths among people with COVID-19 who are not fully vaccinated ranged from to 96.91% in Montana to 99.91% in New Jersey. (Note: Deaths may or may not have been due to COVID-19.)

Source: https://www.kff.org/policy-watch/covid-19-vaccine-breakthrough-cases-data-from-the-states/

Article:

Jul 30, 2021
While information on breakthrough events is still limited and incomplete, this analysis of available state-level data indicates that COVID-19 breakthrough cases, and especially hospitalizations and deaths, among those who are fully vaccinated are rare occurrences in the United States. Moreover, this data indicate the vast majority of reported COVID-19 cases, hospitalizations, and deaths in U.S. are among those who are unvaccinated or not fully vaccinated. These findings echo the abundance of data demonstrating the effectiveness of currently authorized COVID-19 vaccines.

Source: https://www.kff.org/policy-watch/covid-19-vaccine-breakthrough-cases-data-from-the-states/

And there's a visual breakdown of the states with data.

I think, overall, the meme is was accurate.

[UPDATE] However, it's out of date. As I noted below

I think it's clear that 99.2% the meme WAS an accurate figure, but no longer is, for a variety of factors. Mostly because A) more people are vaccinated, and B) highly vulnerable groups (like the 80+ group) have a very high rate of vaccination.

It's also, as other have noted, not a super useful figure, due to the complex inputs.

 

[Agent K]

The 99.2% figure may have been true months ago when fewer people were vaccinated, but it probably doesn't apply to "people currently dying of Covid-19."
The CDC's internal presentation had it at 84.9% in May.
https://www.washingtonpost.com/cont...ections/94390e3a-5e45-44a5-ac40-2744e4e25f2e/

CDC Director Walensky was asked about it at a press conference.

Source: https://youtu.be/26xwZVEOKFU?t=1190

CNN's Kaitlan Collins: Several of you and the President have repeatedly cited figures saying that 99 percent of those who die from COVID-19 are unvaccinated and 95 percent, around that, are hospital- -- who are hospitalized are unvaccinated. With the Delta variant, do you still stand by these numbers, and do you have government data to back them up?

CDC Director Dr. Rochelle Walensky: Yes, thank you for that question, Kaitlan. So those data were data that were from analyses in several states from January through June and didn't reflect the data that we have now from the Delta variant. We are actively working to update those in the context of the Delta variant. I do want to reiterate, though, that based on the data we're seeing -- and we don't have fully updated numbers -- universally, as we look at our hospitalizations and as we look at our deaths, they are overwhelmingly unvaccinated people.

 

[deirdre]

the numbers are changing a bit due to delta it seems (or maybe waning immunity). But everything i read it is still amongst the oldest populations. like here in connecticut we are starting to have issues with it in our nursing homes again because our governor took too long to mandate vaccines for nursing home workers.


either way, there would be a heck of alot more old people dying without the vaccine!

Article:

Of the reported breakthrough cases since February, 53 of those individuals have died, a majority of which were over the age of 75, according to a weekly report on breakthrough cases from the state

...

Despite the rise in breakthrough cases, the state said COVID-19 vaccines do offer protection against the coronavirus. The 53 recorded breakthrough deaths represent 5.8 percent of all COVID-19 deaths since Feb. 9, the state said.

Source: https://www.ctinsider.com/news/article/CT-records-eight-more-COVID-deaths-of-fully-16414839.php

and west virginia

Article:

Officials said at a news conference Monday that 82% of current virus hospitalizations and 74% of the statewide deaths over the past two months involved unvaccinated people.

Source: https://abcnews.go.com/Health/wireStory/latest-zealand-reports-1st-virus-vaccine-death-79718066

 

[JMartJr]

You should of course expect the percentage of those who die to also be vaccinated to increases as the percentage who are vaccinated increases. Obviously, if 100% of the people in a population start off unvaccinated, and you eventually wind up with 100% getting vaccinated, then 100% of the dying would be unvaccinated in the beginning, and 100% would be vaccinated when you are done -- even if there is only one death after everybody in the population under study is vaccinated!

 

[Agent K]

the numbers are changing a bit due to delta it seems (or maybe waning immunity). But everything i read it is still amongst the oldest populations. like here in connecticut we are starting to have issues with it in our nursing homes again because our governor took too long to mandate vaccines for nursing home workers.

Of course if 100% of some group like nursing home residents were vaccinated, then 100% of infected and dying patients in that group would be vaccinated.

 

[LilWabbit]

Percentage of Covid-19 deaths out of all deaths within the two groups ('the fully vaccinated' and 'the not fully faccinated') would perhaps be a more illustrative figure. More importantly, it would not be vulnerable to an entirely opposite (mis)reading once everyone is fully vaccinated, as pointed out by @JMartJr. Figures that merely indicate the distribution of Covid-19 deaths between the two groups is an entirely different metric subject to such a misreading in time.

 

[Mendel]

Article: Published March 1, 2021
Updated April 17, 2021
Former President Donald J. Trump and his wife, Melania, quietly received coronavirus vaccinations in January before leaving the White House, an adviser said on Monday. Source: https://www.nytimes.com/2021/03/01/us/politics/donald-trump-melania-coronavirus-vaccine.html

This is especially notable as Donald J. Trump is a Covid survivor. He had Covid-19 in early October 2020 and was treated at Walter Reed hospital for 3 days.

 

[Agent K]

Percentage of Covid-19 deaths out of all deaths within the two groups ('the fully vaccinated' and 'the not fully faccinated') would perhaps be a more illustrative figure. More importantly, it would not be vulnerable to an entirely opposite (mis)reading once everyone is fully vaccinated, as pointed out by @JMartJr. Figures that merely indicate the distribution of Covid-19 deaths between the two groups is an entirely different metric subject to such a misreading in time.

The normal risk is a useful metric, somewhat invariant to age. I haven't seen it used as a measure of vaccine effectiveness. Poison would lower the relative risk of death from COVID-19 by increasing the risk of death from other causes.

The metric I've seen used for healthcare workers was something like infections per work hour, since more hours generally means more exposure.
Vaccine effectiveness against death is 1 - (vaccinated COVID mortality rate ) / (unvaccinated COVID mortality rate).

Do you want the formula to convert death breakdowns into vaccine effectiveness?

 

[FatPhil]

The 99.2% figure may have been true months ago when fewer people were vaccinated, but it probably doesn't apply to "people currently dying of Covid-19."
The CDC's internal presentation had it at 84.9% in May.
https://www.washingtonpost.com/cont...ections/94390e3a-5e45-44a5-ac40-2744e4e25f2e/

Dying is not hospitalisation. Different stats will have different values, and if part of the efficacy is to reduce the seriousness of the infection, then you'd expect the dying stat to be more extreme than then hospitalisation one.

 

[deirdre]

Dying is not hospitalisation. Different stats will have different values, and if part of the efficacy is to reduce the seriousness of the infection, then you'd expect the dying stat to be more extreme than then hospitalisation one.

the OP meme is still wrong. and the OP data on deaths is dated.

 

[Mendel]

Dying is not hospitalisation. Different stats will have different values,

Obviously. The yellow curve is hospitalizations, the blue curve is deaths, but it excludes those deaths that occurred outside the hospital. For example, patients may die outside the hospital because they had a bad prognosis and were discharged into comfort care.
So that specific statistic is probably only an approximation of the full situation.

 

[FatPhil]

Obviously. The yellow curve is hospitalizations, the blue curve is deaths, but it excludes those deaths that occurred outside the hospital. For example, patients may die outside the hospital because they had a bad prognosis and were discharged into comfort care.
So that specific statistic is probably only an approximation of the full situation.

Thanks for that - I didn't even notice there was a key, and the yellow was unreadable without a loupe even after you pointed it out. Need a bigger monitor...

 

[Rory]

I should really not have been surprised, but I got quite a bit of negative reaction from people saying I was spreading disinformation.

The UAP guys who normally oppose you seemed to like it though.

 

[DavidB66]

For a comparison with the UK, where for several months the Delta variant has been predominant, by the end of June 60% of people dying with the Delta variant had had at least one dose of vaccine, and 43% had had two. There is a useful article from the Guardian here, summarising and explaining the data:

https://www.theguardian.com/theobse...e-who-now-die-with-covid-have-been-vaccinated

As the article points out, Covid deaths are mainly among the elderly, and in the UK the vast majority of elderly people had already been vaccinated at that time. Most of the very elderly (80+), and other groups who are most at risk, had received two doses by the end of March But the vaccines do not give 100% protection, and it is not surprising that many of the deaths are among the vaccinated.

The vaccines seem to be giving only weak protection against infection with the Delta variant. The daily case rate, based on all reported test results, has increased dramatically since its low point in May, thanks to Delta, and is now running at around 30,000 a day. Of course the number of positive test results depends in part on the number of tests, which in turn is affected by factors like the holiday season. A more reliable indicator is the weekly random survey of the population by the Office for National Statistics, using PCR tests. The most recent survey, for week ending 20 August, gave an estimate of over 800,000 cases for the week, implying over 100,000 cases per day. The survey figures have not increased as dramatically as the daily reported case rate.

Despite the increase in infections, the number of hospitalisations and deaths has not increased at the same rate. In very broad-brush terms, infections are running at about 100,000 a day (based on the ONS data), hospitalisations at about 1,000, and deaths at about 100.

The take-up of vaccines has tapered off. Among adults, there is a hard core of people reluctant or resistant to being vaccinated. As a proportion of the total population, the vaccination rate in the UK has now fallen behind that in many other European countries, but this is due largely to the policy on vaccination of children (under-18s). Most other countries allow and encourage vaccination at least from age 12 upwards, but the official advisory body in the UK, the Joint Committee on Vaccination and Immunisation, has been reluctant to recommend this. They take the principled view that they cannot recommend vaccination unless the balance of advantage is clearly in the interests of the individuals concerned. Any benefits to the wider population, in terms of 'herd immunity', are a welcome bonus but not a basis for the decision. In the case of under-18s, where deaths are extremely rare, they think the risks of serious illness or death from Covid are not clearly enough to outweigh the small risks from the vaccine itself. They have recently softened this line to the extent of allowing vaccination for 16 and 17 year olds, on grounds that are not entirely clear.

 

[deirdre]

For a comparison with the UK, where for several months the Delta variant has been predominant, by the end of June 60% of people dying with the Delta variant had had at least one dose of vaccine, and 43% had had two. There is a useful article from the Guardian here, summarising and explaining the data:

https://www.theguardian.com/theobse...e-who-now-die-with-covid-have-been-vaccinated

As the article points out, Covid deaths are mainly among the elderly, and in the UK the vast majority of elderly people had already been vaccinated at that time. Most of the very elderly (80+), and other groups who are most at risk, had received two doses by the end of March But the vaccines do not give 100% protection, and it is not surprising that many of the deaths are among the vaccinated.

which is why people shouldn't be trying to use "death rates" in memes. (esp. wrong death rates)

people should be pushing hospital admission rates and out-of-pocket costs if you end up in the hospital even for a few days with covid.

 

[penk]

As the article points out, Covid deaths are mainly among the elderly, and in the UK the vast majority of elderly people had already been vaccinated at that time. Most of the very elderly (80+), and other groups who are most at risk, had received two doses by the end of March But the vaccines do not give 100% protection, and it is not surprising that many of the deaths are among the vaccinated.

I'm yammering out loud here, and speculating.

I"m wondering if this is a resurgence of the 'normal' mortality that happens regularly with flu-related deaths every year. The last 18 months has seen the population isolated, and extraordinairy measures being taken to prevent transmission of, well, anything.

The elderly population has been vaccinated, and all indications are that if someone does contract COVID via Delta even after vaccineted (a 'breakthrough' infection), the resulting illness is relatively mild (compared to unvaxxed.

For an elderly person, even a mild flu can be absolutely fatal. Are we seeing the numbers coming back up to what would be the norm for coming into the flu season, or at least the people who are dying after vaccinated were people who would, if they got the flu, have a much higher mortality rate in any event?

 

[Mendel]

Are we seeing the numbers coming back up to what would be the norm for coming into the flu season

too early for flu season, I'd venture?

 

[Tedsson]

The elderly population has been vaccinated, and all indications are that if someone does contract COVID via Delta even after vaccineted (a 'breakthrough' infection), the resulting illness is relatively mild (compared to unvaxxed.

That is certainly true.

I have been exceedingly careful over the last 19 months with isolating, masks, hygiene and distancing. First Pfizer vaccination in February, second in May.

Got a positive PCR based diagnosis a few weeks ago. Given my location that is most likely to have been the Delta variant (although they are not specific).

Previously I was a bad candidate for Covid based on age, high blood pressure, size and an as yet undiagnosed cardio-vascular condition.

My "illness" was virtually undetectable. I displayed no real symptoms and had no discomfort much beyond that of having a slight summer cold for a few days. All my signs were good: blood pressure and pulse/blood oxygen fine. The only downside was that, oddly enough, I completely lost my appetite (a serious issue for me!).

Without going all drama queen on it I would probably have died had it not been for the two vaccinations. Thank you Mr Pfizer and Mr BioNTech.

 

[Mick West]

I think it's clear that 99.2% the meme WAS an accurate figure, but no longer is, for a variety of factors. Mostly because A) more people are vaccinated, and B) highly vulnerable groups (like the 80+ group) have a very high rate of vaccination.

It's also, as other have noted, not a super useful figure, due to the complex inputs.

 

[deirdre]

I think it's clear that 99.2% the meme WAS an accurate figure, but no longer is, for a variety of factors. Mostly because A) more people are vaccinated, and B) highly vulnerable groups (like the 80+ group) have a very high rate of vaccination.

It's also, as other have noted, not a super useful figure, due to the complex inputs.

and the military numbers are wrong. even if the meme creator is fudging numbers by using military numbers of today vs. June numbers for death rates.

i'm also a bit suspicious of the physician numbers (from June3-8 ) presented

Article:

A new American Medical Association survey of 301 practicing physicians shows 96 percent are fully vaccinated against COVID-19.

Source: https://www.beckershospitalreview.com/hospital-physician-relationships/most-us-physicians-fully-vaccinated-against-covid-19-small-survey-finds-2.html

just over 1 million physicians, 300 sample size is a margin of error of approx 6%.

As of March 2020, there were just over one million professionally active physicians in the United States

I think it is best to refrain from memes or absolute numbers with Covid, if the goal is to get more people to trust information and get vaccinated.

The real numbers that support vaccination are plenty good, without having to resort to exaggerations that just sow distrust.

 

[bar-bar]

Even with the DoD mandate, I suspect they won't actually hit 100% any time soon. DoD mandated anthrax vaccines for a while in the late '90s early '00s and it took several years to either separate or compel a few holdouts. I'd expect the DoD uniformed services vaccination rate to get above 99% fairly quickly (weeks to a couple of months) and then hover there as the military disciplinary system deals with holdouts.

 

[captancourgette]

Article:

new American Medical Association survey of 301 practicing physicians shows 96 percent are fully vaccinated against COVID-19.

just over 1 million physicians, 300 sample size is a margin of error of approx 6%.

As of March 2020, there were just over one million professionally active physicians in the United States

I assume you're talking about 96% +- 6 (implying to could be more than 100%)
well if you're gonna go down that route then surely most polls are gonna be invalid, eg A poll that puts libertarian support at 2% (margin of error +- 3%)

Personally I would of thought 96% is low, I would of expected at least 99% (caveat: only counting proper medical people and not naturopaths, physic healers etc no matter what their degrees are in these fields )

 

[Mendel]

i'm also a bit suspicious of the physician numbers (from June3-8 ) presented
Article: A new American Medical Association survey of 301 practicing physicians shows 96 percent are fully vaccinated against COVID-19. Source: https://www.beckershospitalreview.com/hospital-physician-relationships/most-us-physicians-fully-vaccinated-against-covid-19-small-survey-finds-2.html

just over 1 million physicians, 300 sample size is a margin of error of approx 6%.

Article:

The survey was fielded from June 3 to June 8, 2021.

The most common reason for not receiving the vaccine was concern about the vaccine being too new and having unknown long-term effects.

• There were no significant differences in physician vaccination rates across various demographic groups, including: PCP vs Specialist, region, gender, age, and race.

• However, there was a significant difference in vaccination rates among Hispanics (84%) and Non-Hispanics (97%).

Source: https://www.ama-assn.org/system/files/2021-06/physician-vaccination-study-topline-report.pdf

It looks like the survey is fairly representative.

Let's see how few physicians could be vaccinated if we're unlucky.
96% means at least 288/301, with 287 it'd be 95%.
If 93.2% or fewer of all physicians are actually not vaccinated, the probability of randomly finding 288 or more who are vaccinated in a survey of 301 physicians is below 5%, so we can be 95% confident that at least 93.3% of physicians in the US are vaccinated.

If we're looking for a double-sided error margin, we can be 95% confident that the real vaccination rate was 92.7%-98.2%.

These numbers are not symmetric (and the 6% error margin is wrong in this case) because the probability is a lot closer to 100% than it is to 50%, which skews the distribution.

(You can check these numbers with https://stattrek.com/online-calculator/binomial.aspx .)

Note that the survey was taken 3 months ago, so the percentage will have gone up by now.

 

[deirdre]

i'm still suspicious. why would only 300 physicians, out of over a million, choose to respond to a poll about vaccines? One might think vaccines to be a more important topic with Delta just starting to hit the states hard. I guess its possible the other million were just busy.

 

[bar-bar]

why would only 300 physicians, out of over a million, choose to respond to a poll about vaccines?

I don't think the poll was sent to all physicians in the US, it seems that this was expected to be a representative sample but I've so far been unable to find an explanation of the poll's methods so it's unclear to me at this time whether this poll was part of a larger one or what the response rate was. It sounds as though this may have been incorporated into a continuous polling operation that WebMD does on behalf of the AMA.

One would hope that the AMA had posted something more thorough than the slide deck Mendel linked to, but if they did I'm not sure where it might be.

Most WebMD general polls (this appears to be distinct from their physician panel) appear to be based upon polls of 1000 of their readers, e.g.

WebMD poll of 1,000 readers, Aug. 24, 2020.

Source: https://www.webmd.com/vaccines/covi...2/webmd-poll-more-acceptance-of-covid-vaccine

I'm unclear if they have a more involved polling operation or not.

In any event, the meme has some sourcing, which is better than many, we're getting into debating the quality of the poll at this point.

 

[Mendel]

i'm still suspicious. why would only 300 physicians, out of over a million, choose to respond to a poll about vaccines?

Methodology:

• AMA developed the survey questionnaire and WebMD programmed and fielded the 5–7 minute survey through their physician panel

They paid for 300 responses (150 each for primary care and specialists), possibly selected to be representative, and when they had the responses in, the system shut down. (It looks like the 301st questionnaire was already being answered when that happened.)

You see that a 300 response survey has a reasonably small error margin for what is essentially a quickly changing number ("The survey was conducted June 3–8 and showed an increase of more than 20 percent for physicians who have been fully vaccinated for COVID-19 compared to a May 2021 Medscape poll.", https://www.ama-assn.org/press-cent...-96-doctors-fully-vaccinated-against-covid-19 ).

The "over a million" you seem to be fixated on is mathematically irrelevant for the quality of the survey if your sample is random enough.
Think of it this way: If you want to find out how much alcohol is in a barrel of whiskey, you only need to draw a small glass. Technically, the 40% alcohol could be on one side of the barrel and the nonalcoholic part on the other, and then you'd get the wrong impression from just a small glass, but practically, it's all well mixed up, so a small sample is good enough.

In this survey, they were ready to check if their doctors had been mixed up well enough, by gender, geography etc., but it turned out that didn't even matter because they didn't really find much of a difference (except for the hispanics). So their "over 1 million doctors" were mixed up so well for the sample that a small number of them sufficed.

Basically, if a small glass is enough to sample the whiskey, it doesn't matter how big the barrel is.

 

[bar-bar]

The "over a million" you seem to be fixated on is mathematically irrelevant for the quality of the survey if your sample is random enough.

I'd caution against this, while your overall point is correct, this statement isn't quite true. You can see that most easily at the limit: a single, perfectly randomly chosen respondent won't provide a meaningful margin of error for even a moderately large population (by analogy, drawing a single, randomly chosen molecule from your imaginary cask would provide a similar result).

I think you may also be hitting the limit of analogical reasoning here, as barrels of whiskey are largely homogenous as a matter of physics and chemistry but population sampling gets more complicated.

I'd say adequacy of sample size is hard to determine without some goal in mind, which I don't think we really have nailed down. Of course, we can say definitively that this poll was deemed by the AMA to be sufficiently accurate for whatever purpose they were using it, and given that we're talking about a meme that seems sufficient to say the meme is reasonably accurate.

 

[deirdre]

we're getting into debating the quality of the poll at this point.

yea. and my original point was "don't use absolute numbers in memes when posting about covid". I do think the physician vax numbers would be fairly high. I just don't think unvaxxed registered WebMD members would go out of their way to advertise the fact. (but that's just opinion).

 

[Mendel]

You can see that most easily at the limit: a single, perfectly randomly chosen respondent won't provide a meaningful margin of error

I didn't say that the sample size is irrelevant, so this point misses the mark.

Article:

• Population size has an impact on the sample variance for small to moderate sized population. For large populations, its impact is minor.

Source: https://www150.statcan.gc.ca/n1/edu/power-pouvoir/ch6/sampling-echantillonage/5214807-eng.htm

I stand by this. I'm ready to write up an example with specific numbers of red and white balls to be drawn at random, if asked.

 

[Agent K]

Some more numbers I came across.

Article:

The W.N.B.A. said in June that 99 percent of its players were vaccinated. The M.L.S. Players Association said in July that it was "approaching 95 percent." This week, the N.F.L. announced it had reached a player vaccination rate of nearly 93 percent. Michele Roberts, executive director of the National Basketball Players Association, said in July that 90 percent of N.B.A. players were vaccinated. This month, the N.H.L. said its player vaccination rate was at 85 percent, and its union warned that unvaccinated players might lose pay if they tested positive.

In tennis, where each player is an independent contractor, there is no player union to encourage unified behavior and no general manager or team owner to encourage vaccination for the team's competitive benefit. Other individual sports are still ahead of tennis, however: The PGA Tour said early this month that its player vaccination rate was "above 70 percent."

Source: https://www.nytimes.com/2021/08/30/sports/tennis/usopen-covid-vaccine-atp-wta.html

 

[Agent K]

The normal risk is a useful metric, somewhat invariant to age. I haven't seen it used as a measure of vaccine effectiveness. Poison would lower the relative risk of death from COVID-19 by increasing the risk of death from other causes.

The metric I've seen used for healthcare workers was something like infections per work hour, since more hours generally means more exposure.
Vaccine effectiveness against death is 1 - (vaccinated COVID mortality rate ) / (unvaccinated COVID mortality rate).

Do you want the formula to convert death breakdowns into vaccine effectiveness?

Here's the formula for vaccine effectiveness (VE):
VE = (1-y/x) / (1-y)
where x= vaccinated share of the population, and y = vaccinated share of patients

To avoid Simpson's paradox, make sure that x and y are from similar groups such as nursing home residents.
For example, if 90% of nursing home residents are vaccinated, but only 50% of nursing home residents hospitalized with COVID-19 are vaccinated, then x=0.9, y=0.5, and VE = (1-5/9) / (1-0.5) = 8/9 = 89%, so the vaccine reduces hospitalizations among nursing home residents by a factor of 9.

 

[Mendel]

Article:

FIGURE 1. Age-adjusted rolling 7-day SARS-CoV-2 infection and hospitalization rates,* by vaccination status† — Los Angeles County, California, May 1–July 25, 2021

* Rolling 7-day incidence was calculated by summing the total number of persons or hospitalizations during a 7-day period and dividing by the total population at the end of the 7-day period.

† Persons were considered fully vaccinated ≥14 days after receipt of the second dose in a 2-dose series (Pfizer-BioNTech or Moderna COVID-19 vaccines) or after 1 dose of the single-dose Janssen (Johnson & Johnson) COVID-19 vaccine; partially vaccinated ≥14 days after receipt of the first dose and <14 days after the second dose in a 2-dose series; and unvaccinated <14 days receipt of the first dose of a 2-dose series or 1 dose of the single-dose vaccine or if no vaccination registry data were available.

Source: https://www.cdc.gov/mmwr/volumes/70/wr/mm7034e5.htm

Case numbers always depend to some extent on who gets tested, but hospitalization rate seems like a very hard statistic.

 

[Mendel]

Data from the UK:

Article:

Getting a COVID-19 vaccine significantly reduces the chance of dying from the coronavirus, real-world data from England suggests.

Figures from the UK's Office of National Statistics (ONS) released Monday found that 0.8% of deaths in fully-vaccinated people were linked to COVID-19 between January and July. These figures covered people who died 21 or more days after the second dose.

For comparison, roughly 37% of deaths in unvaccinated people "involved COVID-19" during the same time period, the data showed.

In total, 57,263 fully vaccinated people in England died at least 21 days after their second vaccine dose, and just 458 deaths "involved" COVID-19. Over the same period, there were 38,964 COVID-19-related deaths in unvaccinated people.

[..]

The ONS data came from census and family doctor health records, considered to be representative of 79% of people aged 10 or older living in England. It didn't specifically look at variants.

Source: https://www.businessinsider.com/covid-vaccine-fully-vaccinated-death-breakthrough-cases-ons-2021-9

38964/(38964+458)=98.8%, this rounds to 99%.
For England, the answer to the question in this forum thread's title is "yes"!

 

[NorCal Dave]

Article:

FIGURE 1. Age-adjusted rolling 7-day SARS-CoV-2 infection and hospitalization rates,* by vaccination status† — Los Angeles County, California, May 1–July 25, 2021

View attachment 46862
* Rolling 7-day incidence was calculated by summing the total number of persons or hospitalizations during a 7-day period and dividing by the total population at the end of the 7-day period.

† Persons were considered fully vaccinated ≥14 days after receipt of the second dose in a 2-dose series (Pfizer-BioNTech or Moderna COVID-19 vaccines) or after 1 dose of the single-dose Janssen (Johnson & Johnson) COVID-19 vaccine; partially vaccinated ≥14 days after receipt of the first dose and <14 days after the second dose in a 2-dose series; and unvaccinated <14 days receipt of the first dose of a 2-dose series or 1 dose of the single-dose vaccine or if no vaccination registry data were available.

Source: https://www.cdc.gov/mmwr/volumes/70/wr/mm7034e5.htm

Case numbers always depend to some extent on who gets tested, but hospitalization rate seems like a very hard statistic.

One would think these 2 charts should settle any questions, especially the hospitalization rate. The vast majority of COVID sick people in the hospital are unvaccinated. No dispute. The same situation is playing out in my much smaller rural county.

AND YET, I still have 2 friends, both in health care, that obsess over the few breakthrough cases, especially ones that result in death. In their minds, and I don't think they're alone in this, if the vaccine isn't 100% perfect, it's "not worth the risk". Very frustrating.

 

[deirdre]

For England, the answer to the question in this forum thread's title is "yes"!

numbers that use figures from January, Feb, March etc are kinda useless if you ask me. because most people were not vaccinated back then during the winter surge.

 

[Mendel]

In their minds, and I don't think they're alone in this, if the vaccine isn't 100% perfect, it's "not worth the risk". Very frustrating.

Not enough people play RPGs. If your RPG character found a potion that reduced the damage you take by even 10%, you'd chug it in an instant.

 

[Agent K]

For England, the answer to the question in this forum thread's title is "yes"!

Afraid not. Among sequenced and genotyped Delta cases, 30% of deaths were among the unvaccinated, and 61% were among the fully vaccinated, partly due to high vaccination coverage of the elderly, and possibly because breakthrough cases are more likely to be sequenced and genotyped.

https://www.gov.uk/government/publi...ars-cov-2-variant-variant-of-concern-20201201
https://assets.publishing.service.g...le/1014926/Technical_Briefing_22_21_09_02.pdf

Anti-vaxxers pointed to the above data, but more relevant data is below.

https://www.gov.uk/government/publications/covid-19-vaccine-surveillance-report
https://assets.publishing.service.g.../Vaccine_surveillance_report_-_week_37_v2.pdf

The plot of reported cases looks weirder.

 

[DavidB66]

I think the data quoted above for Covid deaths in England are based on this study by the Office for National Statistics.

https://www.ons.gov.uk/peoplepopula...nd/deathsoccurringbetween2januaryand2july2021

Press reports have tended to highlight the fact that in this data only a very small percentage of deaths (~1%) occurred in people who had received two doses of vaccine, and of these only about half were people infected 21 days or more after the second vaccination.

As has been pointed out above, this is not as impressive as it may seem at first sight, since it covers only the first 6 months of the year, when comparatively few people in England had been 'fully vaccinated' as defined in the study. Speaking for myself, I had my second jab in mid-April, and allowing a 21-day gap I was not 'fully vaccinated' until mid-May. On the other hand, I am not in the very highest risk groups by age and medical history, and these groups were given top priority for vaccination. Many of the highest risk people, who account for the bulk of Covid deaths, will have been 'fully vaccinated' by the end of March.

The ONS study does also give 'age-adjusted mortality rates' (see section 4 of the article). I take it that these are intended to allow for the fact that older people are not only more likely to die of Covid but (in the UK) more likely to have been vaccinated at any given date during this period. But I haven't dug down into the technical details, and probably wouldn't understand them anyway. The article itself mentions some caveats in the last para of section 4. Taking the graphs in Figure 1 at face value, they do seem to confirm that at all times people who had received a second dose of vaccine were both relatively and absolutely well-protected against death from Covid. For the first half of the period even one dose seems to have given good protection compared with the unvaccinated. For the second half the picture is not clear; it even seems that those with a single dose were more likely to die than the unvaccinated, but perhaps the highest-risk people among the unvaccinated had already paid the price by dying during the peak mortality of January to March.

A further big caveat is that the study is too early to reflect the full impact of the Delta variant since June, and the effects of relaxing many of the 'rules'. The daily reported new cases (positive tests) have increased from about 3,000 a day to 30,000 a day since the beginning of June. Hospitalisations and deaths have also increased, but not nearly to the same extent, which is attributed to the high levels of immunity due to vaccination and/or prior infection. Compare the graphs for 'daily new cases' and 'daily new deaths' in the Worldometers page for the UK to see the contrast.

https://www.worldometers.info/coronavirus/country/uk/

 

[Cryptic]

How can 99% of deaths be unvaccinated with information like below publicly available?

"Vermont pretty much blew Fauci's data out of the water for September. "Just eight of the 33 Vermonters who died of Covid-19 in September were unvaccinated, the Vermont Department of Heath said Wednesday." Take a look:"
https://vermontdailychronicle.com/2021/09/30/76-of-september-covid-19-deaths-are-vaxxed-breakthroughs/

Vermont — 76% of September Covid deaths were Fully Vaccinated…

https://citizenfreepress.com/breaking/vermont-76-of-september-covid-deaths-were-fully-vaccinated/

 

[Landru]

How can 99% of deaths be unvaccinated with information like below publicly available?

"Vermont pretty much blew Fauci's data out of the water for September. "Just eight of the 33 Vermonters who died of Covid-19 in September were unvaccinated, the Vermont Department of Heath said Wednesday." Take a look:"
https://vermontdailychronicle.com/2021/09/30/76-of-september-covid-19-deaths-are-vaxxed-breakthroughs/

Vermont — 76% of September Covid deaths were Fully Vaccinated…

https://citizenfreepress.com/breaking/vermont-76-of-september-covid-deaths-were-fully-vaccinated/

I would contend that the sample size of Vermont is too small. https://en.wikipedia.org/wiki/Sample_size_determination?wprov=sfla1

A common problem faced by statisticians is calculating the sample size required to yield a certain power for a test, given a predetermined Type I error rate α.