This year we kill it: Corona Virus Thread Mark II

[Tvtyrant]

Ok, I have a few questions about the booster shot:

1. Do I get sick from the booster shot?

2. It is even necessary to get a booster shot even if I'm fully vaccinated?

1. A booster shot is an add on to keep your body ready to fight covid. It will probably make you slightly more ill then the third shot, since your body reacts more strongly each time (this is why food allergies tend to become apparent after an initial meal that didn't seem adverse.)

2. It might not be, no one is sure yet for non-elderly individuals whether it will be important or not. If Covid keeps spinning off variants then it likely will be, yes.

[Tyndmyr]

Depends on what they mean by 86.3%, if they are taking the death rate due to unvaccinated covid as 100% then reduction of that to 13.7% is pretty good if not ideal, if on the other hand they are taking the whole population as 100% that's pretty near to the unvaccinated death rate.

Still quite useful, yes. But definitely quite far from "100% effective in preventing serious illness" yknow? Huge gap between those.

A 20-year-old girl died this week over here because of a stroke, after being vaccinated with Janssen vaccine. It has not been confirmed that the vaccine itself has been the cause of her death, but that hasn't stopped the anti-vax crowd of grabbing it and running with it.

That's hard to confirm conclusively. You can rule out other causes, I guess, but if she didn't have any condition and nothing obvious shows up on investigation, all you really have to go on is the timing, which can never be entirely certain.

That said, healthy 20 year olds rarely get strokes in the first place, and it is *quite* rare for a healthy 20 year old to die as a result.

Of course, just because someone else had a side effect doesn't make it necessarily dangerous for you. It may be that there is some interaction with a specific health condition. A lot of the rarer side effects require somewhat specific conditions, and if that doesn't fit your circumstances, it may not matter.

That is quite difficult to explain to a grieving father, though. Isn't going to matter much to him if other people were fine. About all one can do is offer sympathy.

Yes and that also plays into the trust deficit we are currently seeing.

Yeah, there's a whole lot of that goin' around. Don't think there's any quick fix, either. Trust is earned, and slowly....but a few bad actions can damage that trust relatively quickly.

Be it agencies, pharma companies, or whatever, existing distrust from past misdeeds definitely play a part in how people act today. Can't even really blame them for it. Sure, some of the more conspiratorial stuff is a little bonkers, but that stuff is relatively rare, and even that is often connected to past fears, albeit in a fairly non logical way.

Things like scientific credibility get ignored as not being all *that* critical...until something like this happens, and then it is. Not everyone's an expert, and there is inherently some uncertainty in the situation, so trust becomes essential.

This whole thing has at least somewhat altered my views on how humans react to disaster/how history happens. It's easy to mock the zombie movie for people doing dumb stuff, but in an actual bad scenario, emotions and trust matter immensely.

[Bartmanhomer]

1. A booster shot is an add on to keep your body ready to fight covid. It will probably make you slightly more ill then the third shot, since your body reacts more strongly each time (this is why food allergies tend to become apparent after an initial meal that didn't seem adverse.)

2. It might not be, no one is sure yet for non-elderly individuals whether it will be important or not. If Covid keeps spinning off variants then it likely will be, yes.

Ok. Thank you for telling me about this information. In all honesty, this pandemic should have been over already and it just one annoying TV episode after another episode.

[NichG]

Yeah, there's a whole lot of that goin' around. Don't think there's any quick fix, either. Trust is earned, and slowly....but a few bad actions can damage that trust relatively quickly.

Be it agencies, pharma companies, or whatever, existing distrust from past misdeeds definitely play a part in how people act today. Can't even really blame them for it. Sure, some of the more conspiratorial stuff is a little bonkers, but that stuff is relatively rare, and even that is often connected to past fears, albeit in a fairly non logical way.

I think the productive blame, such as it is, is centered around a particular fallacy that people fall prey to, which I hold people responsible for learning to think their way around and avoid in the future: namely, if someone you don't trust says something, you shouldn't actively find your belief in the opposite confirmed by that - at most you should just not use them as a source of evidence. It's similar to the idea that if someone says something you agree with strongly and something you disagree with, and other people attack the person as a whole, you might get baited into defending the thing you disagreed with because you're seeing everything the person says as if it has to be taken together. Similarly, if you give someone you distrust the power to make you believe the opposite, that's just the same as trusting them.

[warty goblin]

Remember, something on the order of half of all humans on the planet have had at least one covid vaccine in the last year. Just by dint of numbers you've got a very high probability of seeing events that are, for any given person, extremely unlikely because you're talking about a population of 3.5 billion people. This for instance estimates that about .2% of the female (US) population ages 20 - 29 will suffer a stroke in a given year, which a rare event but over the hundreds of millions of ~20 year old women who have had a covid vaccine, the probability of one of them suffering a stroke in the next few days becomes quite high.

Back-of-the-envelope probability, let's assume a 1% case fatality rate for strokes in 20-29 year old women, so we would expect about .002% of the population to die in a given year, or roughly .00004% to die per week, for a final per week fatality probability of .0000004. This is very small, you shouldn't worry about otherwise healthy young women you know dying of stroke. But we are talking about large populations, where the math behaves differently. If a hundred million 20 - 29 year old women are vaccinated, and the vaccine has no impact on stroke risk, the probability of none of them dying in the next week is (1 - .0000004)^(100000000) = 4x10^(-18), a number that is for all practical purposes is zero. Even if only ten million such people were vaccinated, the probability of no deaths is under 2%.

[halfeye]

The last I heard, which was a while ago, the death rate for strokes was 1/3rd, however, that means the crippled rate for strokes is 2/3rds, because if it clears up quickly it's not a stroke, it's a TIA.

[Tyndmyr]

Looks like the forum ate my previous post, trying again, albeit more tersely.

Ok, I have a few questions about the booster shot:

1. Do I get sick from the booster shot?

It's the same stuff as the first two shots. Your reaction will likely be similar to the reaction to the second shot.

2. It is even necessary to get a booster shot even if I'm fully vaccinated?

This is currently disputed. The NIH panel studying it concluded that they were not generally needed. CDC head overruled them and recommended it for some.

I would suggest that if you're pondering it, consult your personal doctor, and talk with them about concerns you have either way. They have your medical history, and can certainly make a better determination than I can. We can chat in generalities, but doctors are best for specifics.

Remember, something on the order of half of all humans on the planet have had at least one covid vaccine in the last year. Just by dint of numbers you've got a very high probability of seeing events that are, for any given person, extremely unlikely because you're talking about a population of 3.5 billion people. This for instance estimates that about .2% of the female (US) population ages 20 - 29 will suffer a stroke in a given year, which a rare event but over the hundreds of millions of ~20 year old women who have had a covid vaccine, the probability of one of them suffering a stroke in the next few days becomes quite high.

Back-of-the-envelope probability, let's assume a 1% case fatality rate for strokes in 20-29 year old women, so we would expect about .002% of the population to die in a given year, or roughly .00004% to die per week, for a final per week fatality probability of .0000004. This is very small, you shouldn't worry about otherwise healthy young women you know dying of stroke. But we are talking about large populations, where the math behaves differently. If a hundred million 20 - 29 year old women are vaccinated, and the vaccine has no impact on stroke risk, the probability of none of them dying in the next week is (1 - .0000004)^(100000000) = 4x10^(-18), a number that is for all practical purposes is zero. Even if only ten million such people were vaccinated, the probability of no deaths is under 2%.

There are several statistical errors here. There is no valid population of 3.5 billion people that are comparable. The event also happened in Slovenia, so the population of the US isn't relevant. The population of Slovenia is about 6 million, and there are approximately 60,000 women within a couple years of her age, forming a substantially similar cohort.

Next up, your risks are off. Risks increase exponentially for strokes with age, so the risk of a 20 year old woman having one are substantially lower than the average risk of the 20-29 cohort. Risk of stroke is primarily determined by age, smoking, heavy drinking, cholesterol and blood pressure. For instance, a heavy smoker(20 cigarettes/day) is at a six times greater risk of stroke than a non smoker who is otherwise identical. It is extremely rare for a stroke to happen with none of these risk factors at such an early age.

No risk factors were reported for this woman.

Furthermore, risk of death to stroke, within one month, is one in eight. Some people do die later, with notable statistical differences up to a year later, but this is not relevant to this case. She got the vaccine, then a stroke, and was dead in two weeks. Anyone who dies a year after the shot would...literally not have died yet, and thus cannot be a relevant part of this comparison. We are specifically looking at the rarer relatively sudden deaths.

Given all that, the odds of this happening due to an unrelated accident appear to be rather more unlikely than a 1/100 shot. The decision to suspend use of the vaccine while they investigate further appears reasonable.

That isn't the same thing as saying the vaccine is "bad"...it may turn out that there is some other reason that was overlooked. Or perhaps the vaccine is dangerous, but only in extremely limited circumstances that we can avoid. Investigating further is a pretty reasonable outcome.

[pendell]

In other news, new covid cases are down 25% .

But there are other issues. There is a somewhat disturbing long-term symptom . Sort of like restless leg syndrome but it isn't your LEG that won't hold still. Um.

Whatever. So long as it dies (or at least recedes to an endemic disease we get boosters for every year without overloading our hospitals) I'll be satisfied.

Respectfully,

Brian P.

[Ibrinar]

@warty goblin 2 per thousand huh didn't think it was that many in that age range. I mean that would still be 2 per 100k with your assumed 1% death rate and at 20 this says about https://www.ssa.gov/oact/STATS/table4c6.html 0.000405*100000= 40 woman per 100000 die per year. Which would put it at 2% of the deaths. Hmm ah https://www.worldlifeexpectancy.com/...age-and-gender for woman in the 15-24 age range you get 65 strokes out of 6620 deaths so about 1%. (us data) For 25-34 it is 276/14565=1.9%. Guess your estimate is pretty close. (1/(1%*0.000405)=> 1 in 247k . 274k*52=> 1 in 14,248,000 (to restrict to a week) is I think easier to parse as estimate)

There are several statistical errors here. There is no valid population of 3.5 billion people that are comparable. The event also happened in Slovenia, so the population of the US isn't relevant. The population of Slovenia is about 6 million, and there are approximately 60,000 women within a couple years of her age, forming a substantially similar cohort.

Unless the country is important restricting it to slovenia is arbitrary and by itself a stat error because you might have heard about the event if it happened it many other countries. (That they aren't entirely comparable is not sufficient reason to only look at an arbitrary subset.) Let's say an event happens to 1 person out of a group of 1 million, you consider the statistics of the situation and come to some result. I then tell you that I have split the group arbitrarily into 10 smaller groups, does that change your results? Now I declare I split them based on city for instance, and that might change things but unless I have reason to believe the probabilities between the cities are vastly different it still shouldn't be treated as 1 event per subgroup size in question since that would warp the numbers. Because if it was a 1 in 500k event with the same probability in each subgroup that would make it sound rare while chances were there would be one among the million.

I mean otherwise we might as well restrict it to whatever city she lived in to make it even more surprising. (On the other hand it is also pointless to analyze the statistics of this case without checking how many strokes have happened close to a vaccination overall. edit- removed a sentence I started but forgot to finish and can't be bothered to finish now)

[NichG]

Talking about the statistics of a single event or small integer number of events in particular is vulnerable to that kind of distortion as well...

[Tyndmyr]

Unless the country is important restricting it to slovenia is arbitrary and by itself a stat error because you might have heard about the event if it happened it many other countries. (That they aren't entirely comparable is not sufficient reason to only look at an arbitrary subset.) Let's say an event happens to 1 person out of a group of 1 million, you consider the statistics of the situation and come to some result. I then tell you that I have split the group arbitrarily into 10 smaller groups, does that change your results? Now I declare I split them based on city for instance, and that might change things but unless I have reason to believe the probabilities between the cities are vastly different it still shouldn't be treated as 1 event per subgroup size in question since that would warp the numbers. Because if it was a 1 in 500k event with the same probability in each subgroup that would make it sound rare while chances were there would be one among the million.

I mean otherwise we might as well restrict it to whatever city she lived in to make it even more surprising. (On the other hand it is also pointless to analyze the statistics of this case without checking how many strokes have happened close to a vaccination overall. Side note restricting it to a narrow couple years age range also causes problems, yes the change of probability has to be taken into account)

Restricting by country is unlikely to be arbitrary, as each country has its own vaccination plan, using different forms of vaccinations. It wouldn't be reasonable to treat all different vaccinations as the same thing. There does not appear to be different vaccines used on a city by city level, so that level would indeed be arbitrary.

Furthermore, if we look at other countries, we must remember that this is not the only reported stroke incident attributed to the covid 19 vaccination.
Initial AstraZenica symptoms, per Cambridge.
Heart.org warning of six additional strokes in young women taking specifically the J&J vaccine

If you were doing a multi-country analysis, you'd need to consider all the different types of vaccinations(including boosters, where those are performed) and all the stroke reports. This would be, well, probably a great deal more complicated than most forum posts.

However, even cursory investigation quickly disproves the idea that this is a one off event, and any reasonable multi-country analysis with that many reports over J&J's average usage is...unusually high.

Further investigation is definitely warranted.

[sktarq]

Honestly wouldn't such circulatory events be absolutely expected?

We know that the spike protein of the SARS-COV-2 can have effects that involve the circulatory system, especially causing thrombosis.

This, from what I've gathered, is because there is a degree of similarity in shape to a chemical in the later stages of the waterfall that causes thrombosis/scab formation.

So spike proteins carried by the virus can cause blood clotting issues.

vaccines work by providing spike proteins and irritants to get the immune system to recognize the spike protein as a threat so that when it appears with the virus attached the host immune system will attack it forthwith.

So it involves a fair amount of spike protein circulating in the blood.

So it would logically follow that vaccines would be a risk factor for blood clotting issues.

Now immunology is where logic goes to die, but even so this would be a pretty expected side effect to see in extreme cases.


Now the frequency of such an event matters a lot. You are going to kill people by giving them a vaccine, by not giving them a vaccine, by ordering half of them to get the vaccine, you will kill people by ordering them to stay home (via depression, knock on effects of job loss, and domestic violence to start), you will kill people by ordering them to stop staying home (via totally unrelated disease exposure, transport accidents, and walking down the wrong alley)....hell you'd almost certainly kill people if you asked 100K Londoners to drive to John O Groats and back. Some poor sod would likely get in accident on the A1 Motorway and die.

We are going to have to expect something like these events to trickle through for any and all courses of action. Just like we accept the benefits of traveling at over 20mph will kill people we will have to accept that ending lockdowns or vaccines will kill people, but we think as a society that the cost benefit analysis is worth it.

[Ibrinar]

Restricting by country is unlikely to be arbitrary, as each country has its own vaccination plan, using different forms of vaccinations. It wouldn't be reasonable to treat all different vaccinations as the same thing. There does not appear to be different vaccines used on a city by city level, so that level would indeed be arbitrary.

For somewhat rare events where you get at least a few dozen or something in the region it makes sense to just separate data sets because the numbers are high enough that the local dataset should be close enough to the probabilities to look at local ones. However for one off events where you look whether it happening randomly is plausible separating the dataset like that gives you a nonsense answer for how unlikely it is. And warty goblin was analyzing it as one. Of course the stroke wasn't a one off event so analyzing it like that is a bit pointless but if you do restricting it to the country is the wrong way to go.

[Yakk]

J&J/AZ vaccines are traditional viral vector vaccines.

They infect you with a wimpy virus that hijacks the DNA of some cells and gets it to pump out Covid-19 virus proteins (in particular, the spike).

Viruses are mediocre about hacking our DNA. They can do it wrong sometimes, and you'll get mutant spikes produced. Those mutant spikes may be why people are rarely getting those clotting problems with the J&J and AZ viral vector vaccine.

So for Viral Vector:
* Crippled virus infects cell.
* Crippled virus hacks that cell's DNA.
* Infected cell produces mRNA instructions to make Covid-19 spike protein.
* Cell follows mRNA instructions, produces spike protein.
* Human immune system sees alien proteins being made by a human cell, attacks
* Immune response develops

The crippled virus can't spread, because it is crippled. It is also based off a virus that can barely infect humans to start with (I think AZ used a chimp virus).

The mRNA vaccines, instead of hijacking your DNA with a virus, we just slide in the mRNA:
* mRNA contained in microscopic bubbles of fat (lipids) slips into human cell.
* Cell follows the mRNA instructions, produces spike protein.
* Human immune system sees alien proteins being made by a human cell, attacks
* Immune response develops

The lack of having to hack the DNA and produce the mRNA inside the cell may be the reason why the mRNA viruses are less prone to that error.

The other problem with the viral vector vaccines is that the human gets immune to the viral vector. And can kill/suppress the viral vector before it managed to produce a significant amount of spike protein, especially in the 2nd injection.

This may be why cross-vaccination studies have found that AZ followed by an mRNA vaccine produces immune response a lot like a double-mRNA vaccination. The second AZ dose runs into viral vector immunity, which makes it less effective at generating Covid-19 immunity.

Anyhow, that is my current understanding.

This is also interesting in that Novavax is getting approval for a vaccine that just contains pre-printed Spike proteins.

[warmachine]

Adenovirus vectors don't alter host DNA as they lack reverse transcriptase. You're thinking of retroviruses. Adenovirus vectors contain DNA but this is merely inserted into the cell nucleus, not integrated into the host's DNA. Better to say they hack the cell's protein production.

[Tyndmyr]

Honestly wouldn't such circulatory events be absolutely expected?

...

So it would logically follow that vaccines would be a risk factor for blood clotting issues.

Now immunology is where logic goes to die, but even so this would be a pretty expected side effect to see in extreme cases.

That is certainly possible, but it's...very complicated indeed. How much clotting is dangerous can vary a great deal depending on other factors.

Now the frequency of such an event matters a lot. You are going to kill people by giving them a vaccine, by not giving them a vaccine, by ordering half of them to get the vaccine, you will kill people by ordering them to stay home (via depression, knock on effects of job loss, and domestic violence to start), you will kill people by ordering them to stop staying home (via totally unrelated disease exposure, transport accidents, and walking down the wrong alley)....hell you'd almost certainly kill people if you asked 100K Londoners to drive to John O Groats and back. Some poor sod would likely get in accident on the A1 Motorway and die.

Indeed, but the math is naturally complicated by a great deal of uncertainty*, as well as the fact that the risk isn't evenly distributed. It is likely that these sorts of events are only a risk factor for some, while others are in essentially no danger. We already know that there's a significant gender and age skew, but there could be a lot more information. Ideally, we'd want to know what, allowing us to use it where there is no risk, and avoid it where risk is substantial.

Observing a problem doesn't mean that the correct answer is to abandon the vaccine forever, and that's not even what's being proposed here. It's just a suspension of that type of vaccine while further study is being done. This is one of the advantages of having multiple vaccines, we can swap around as we learn more.

We are going to have to expect something like these events to trickle through for any and all courses of action. Just like we accept the benefits of traveling at over 20mph will kill people we will have to accept that ending lockdowns or vaccines will kill people, but we think as a society that the cost benefit analysis is worth it.

Certainly people will unavoidably die no matter what. However, I think that, ethically, can get thorny. We don't want to get too dismissive of deaths just because we can't save everyone. We also should be careful of forcing unequal risk on people. For instance, a kidney transplant recipient was recently denied by the hospital because she wasn't vaccinated. Given that she has end stage renal failure, well...living without kidneys isn't really in the cards. I'm not certain that this decision made her, or anyone else, safer.

Especially because vaccination/etc decisions can be really complex for people in those circumstances. They are usually at highly elevated risk from side effects because they are not at all healthy. Flip side, covid is probably really awful for them as well for similar reasons.

We all accept that some degree of random risk arises when we drive down the highway, but that's very different from specifically imposing risk or harm on specific people.

*I'm assuming the folks making these vaccines likely know a great deal more about the topic than I, and certainly they are unable to predict every possible issue. This is particularly true for pandemics when time is a big factor, and we're working off limited knowledge of the disease itself. To some extent, we need to take some risk, because the cost of not doing so can be high. But that risk doesn't vanish, and as we learn more, well, we need to adjust.

[farothel]

According to studies the risk of thrombosis from the vaccines is lower than from COVID itself: https://www.news-medical.net/news/20...ccination.aspx

This only talks about AstraZeneca and Pfizer vaccines, but for J&J the risks are comparable with AstraZeneca. And now the risks are known, people who have symptoms of thrombosis and who indicate that they have recently been vaccinated will be checked for that and treated where needed.

[Bartmanhomer]

Well, my older brother got his second vaccine today. I'm very proud of him.

[pendell]

If you've got J&J vaccine, strongly recommend a booster . It's not that we're seeing tons of breakthrough infections or what not, so there's still a degree of safety, but as Dr. Fauci says in the clip, this really should have been a 2-dose vaccine from the beginning.

Respectfully,

Brian P.

[sihnfahl]

For instance, a kidney transplant recipient was recently denied by the hospital because she wasn't vaccinated. Given that she has end stage renal failure, well...living without kidneys isn't really in the cards. I'm not certain that this decision made her, or anyone else, safer.

I think this needs to be added. While she DOES have end-stage kidney disease, which requires regular dialysis, the issue with transplants is that you need to take immunosuppressants to fight off tissue rejection. Barring cloning, it's still a part of someone elses' body that's in them, and their body sees it as an invader, no matter the function of the organ.

Transplant recipients have a higher rate of negative outcome of COVID due to the immunosuppressants hindering the body's ability to mount a defense.

Not to mention have a higher chance of contracting COVID after exposure to the virus because of the weakened immune system.

That's why they're insisting before the procedure. So that the immune system has a chance to learn before they weaken it.

[Tyndmyr]

Not to mention have a higher chance of contracting COVID after exposure to the virus because of the weakened immune system.

That's why they're insisting before the procedure. So that the immune system has a chance to learn before they weaken it.

Largely irrelevant thanks to the vaccine not effectively preventing illness.

Or, after some period of time, even at preventing death. Here in MD, most recent stats have 40% of recent covid deaths among the fully vaccinated. This is the ugly fact behind why boosters are being pushed...we're seeing results vastly below what was originally postulated for the vaccines.

It's not no effect, but it's...not great. Especially over the long term. So someone taking a transplant is going to fairly rapidly be in that boat regardless of what shot they take now. It's a bit of a rough scenario regardless of their choice on the shot. Other infection avoidance measures probably become critical for them.

[georgie_leech]

Worth noting that [most[/I] of the population being vaccinated but being less than 50% of the deaths is actually strong evidence of it preventing serious cases.

[sihnfahl]

Or, after some period of time, even at preventing death. Here in MD, most recent stats have 40% of recent covid deaths among the fully vaccinated. This is the ugly fact behind why boosters are being pushed...we're seeing results vastly below what was originally postulated for the vaccines.

Under 30%, past six weeks, and it ignored comorbidities.

"The state department of health agrees, writing in a statement, "As the number of our residents who are vaccinated continues to increase, we expect to see an increase in the proportion of COVID-related deaths occurring in vaccinated individuals. Many of these deaths are linked to comorbidities that make patients more vulnerable.""

Colin Powell. He was vaccinated, he still died. Would you call that a failure of the vaccine, or the fact he was an elderly man with cancer that had a higher risk DESPITE being vaccinated?

And the boosters are because of the Delta variant. Against the original strains, the vaccine is still very effective; it loses some of that effectiveness against the new strains, so the boosters help with that.

[NichG]

Better to talk numbers instead of 'good' or 'bad' protection.

About 66% of Maryland is double-vaxxed. However there's an extra 10% single vaxxed and about 10% have had COVID (so conservatively assuming no correlation between having had COVID and vax status, that's still another 4%). So the control population constitutes 44% of the total population and 14% of the total population is part of that control and has some immunity already - e.g. about a third.

So if 30% of deaths are breakthrough, the raw risk reduction of being in the 2x group as opposed to the rest is a bit over half. However, if we take that a third of that group also have protection, we have for relative risk vs naive of x:

0.66x / (0.66x + 0.34(2/3 + x/3)) = 0.3

X = 0.159

Or, compared to no protection it's about 84%. Which is less than the original studies showed by a factor of 2 to 3 - not great news, but not as bad as the ~60% effectiveness you'd estimate from treating everyone not double-vaxxed as a naive population. It's about 2-3 times the protective effect of people around you cloth masking, or a bit less effective than everyone around you wearing N95s (~90%). It should likely stack multiplicatively with that, and doesn't really require any thought or effort once you have it.

If a booster takes that 84% back up to 95% or more, great! That's another factor of 3 multiplier in protection.

A note about assumptions in this - if there are more cases of COVID than reported, it makes the vaccine effectiveness vs naive higher. The less effective single-vaxing or previous COVID is as compared to double, the less effective that makes the vaccine in the analysis. I don't have great estimates for either of those. Additionally, waning immunity and demographic effects can matter a lot here. If e.g. there is a strong correlation between age and vax status, you'll get Simpson's paradox effects (because the older you are, you're much more likely for any COVID case to be fatal by orders of magnitude). Imagine e.g. the vaccine provided 99% protection but COVID was 1000x as lethal in 65+ as in 65-. If the vax fraction in 65+ were 10% higher than in 65-, the naive calculation would show the vaccine as a wash.

[Tyndmyr]

Worth noting that [most[/I] of the population being vaccinated but being less than 50% of the deaths is actually strong evidence of it preventing serious cases.

It is consistent with the vaccine still having *some* effect, but fairly mildly so. It reduces incidences of serious cases, it is certainly not protecting against them entirely, or even very frequently. After all, death is about as serious as it gets.

So long as the fully vaccinated rate is higher than the death rate in fully vaccinated people, that implies they are dying less frequently. But not much less.

Colin Powell. He was vaccinated, he still died. Would you call that a failure of the vaccine, or the fact he was an elderly man with cancer that had a higher risk DESPITE being vaccinated?

Comorbidities exist in both vaccinated and unvaccinated people, and are thus not a delta relevant to the vaccine's overall efficacy. Overall covid deaths of all sorts have been strongly linked with comorbidities.

And the boosters are because of the Delta variant. Against the original strains, the vaccine is still very effective; it loses some of that effectiveness against the new strains, so the boosters help with that.

The boosters are the same material as the original, there isn't yet a delta specific dose. For that matter, there isn't a delta specific field test, so how much of illness is due to delta is subject to more error. They *can* differentiate in a lab, but the vast majority of tests are not this.

In any case, discussing how well the vaccine would have worked if no changes had taken place is largely a moot point. The variants are here, and enough have developed that we can reasonably expect more variants to continue to develop. The vaccine is, right now, far less effective than originally hoped.

Better to talk numbers instead of 'good' or 'bad' protection.

About 66% of Maryland is double-vaxxed. However there's an extra 10% single vaxxed and about 10% have had COVID (so conservatively assuming no correlation between having had COVID and vax status, that's still another 4%). So the control population constitutes 44% of the total population and 14% of the total population is part of that control and has some immunity already - e.g. about a third.

Having had covid, but not the vaccine, will not result in the illness being reported as a breakthrough case, but as a case among the unvaccinated. There is relatively little tracking of who is resistant due to having had covid, with a great deal more reporting on vaccination status.

Therefore, this whole bit isn't consistent with the way numbers are reported, and would need to be cut.

You also have too high of numbers for single shot. 66% of MD has had 2+ shots. Another 7% has had only a single shot. Stats from usafacts.org.

The reported breakthrough rate was *only* among those who had already had two shots.

Lastly, these vaccination rates are from *now* not from back when the infections happened. Vaccination rate has risen over time.Each of these factors results in a lower amount of protection than you calculated.

As for unreported cases, well...there is little we can do about data we don't have, it'd be pure speculation. Given that a *lot* of companies mandate weekly testing for all unvaccinated here, it seems highly probable that this number is fairly low. If any infections are being missed, it would more probably be among the vaccinated, who do not have this testing mandate. However, we have no good way to predict how large that effect is.

As a footnote, we also have over 15% of the MD population with a booster. This hasn't been the case long enough to have good data on it yet, but if there is a significant infection risk variance between 2 and 3 shots, that might also be a complicating factor.

Edit: Realized I didn't properly summarize. What this largely boils down to is a decent argument for waning protection against severe illness. Earlier time periods had remarkably lower death rates among the vaccinated, with one period being heavily advertised as wholly unvaccinated deaths. This appears to be significantly diminishing as time passes. While breakthrough infections were reasonably common even in the earlier timeframe, deaths were quite a bit rarer. We're sort of using death as a proxy for severe illness...but it's the outcome that is a fairly concrete metric and it's pretty objective, and preventing it is arguably the most important, or at least one of the most important purposes of a vaccine.

This probably means boosters will be tried. How well that works, well...that depends on a lot. Variants(some of which might not yet exist) are one potential wild card. We may also need to start redesigning vaccines. Either to account for variant specific changes, or to incorporate more viral features to make the vaccine less vulnerable to small variances. Natural resistance should be helping us along the way, though, so even if there's another spike, hopefully the overall intensity continues to recede.

[sihnfahl]

Comorbidities exist in both vaccinated and unvaccinated people, and are thus not a delta relevant to the vaccine's overall efficacy. Overall covid deaths of all sorts have been strongly linked with comorbidities

Which goes back to the vaccination for transplant patients.

"Clinical effectiveness of COVID-19 vaccination in solid organ transplant recipients"

"Among 2151 SOTRs, 912 were fully vaccinated, and 1239 were controls (1151 unvaccinated, 88 partially vaccinated). Almost 70% of vaccinated subjects received the mRNA-1273 vaccine. There were 65 cases of COVID-19 that occurred during the study period – four occurred among fully vaccinated individuals and 61 among controls (including two in partially vaccinated individuals). Incidence rate for COVID-19 was 0.065 (95% CI 0.024–0.17) per 1000 person days in vaccinated versus 0.34 (95% CI 0.26–0.44) per 1000/person days in the control group; IRR was 0.19 (95% CI 0.049 −0.503, p < 0.005). There were no COVID-19 related deaths in the four breakthrough infections and two of 61 (3.3%) among controls.

Conclusion
We demonstrate real world clinical effectiveness of COVID-19 vaccination in SOTRs with an almost 80% reduction in the incidence of symptomatic COVID-19 versus unvaccinated SOTRs during the same time."

80% reduction in incidence.

That is not an insignificant level.

[NichG]

Having had covid, but not the vaccine, will not result in the illness being reported as a breakthrough case, but as a case among the unvaccinated.

Yes, this is the point. If we're talking about the percentage of cases which are breakthrough, then that's the ratio of: (Cases in 2xVax)/(Cases in 2xVax + Other Cases)

So if there is some immunity in the Other Cases category, it will make the vaccine appear less effective than it actually is.

For example, lets say a naive population of some reference size would have had N cases of COVID during a monitoring period, and we had the same number of 2xVax as non-2xVax for simplicity in this example. If the true effectiveness of the vaccine vs naive were 90% (meaning that 90% of cases you would have seen don't happen), then the apparent effectiveness given X degree of immunity present in the Other Cases category would be: (N*0.1)/(N*0.1 + N*(1-X))

If X=0 (no latent immunity in the 'other' category), you'd end up with 9.1% of cases being breakthrough cases. If X=0.5 (50% latent immunity in the 'other' category), you end up with 16.7% of the cases being in the 2xVax group. If X=0.9 (everyone who hasn't been vaccinated has had Covid and the degrees of immunity are the same), you get 50% of the cases being in the 2xVax group.

You also have too high of numbers for single shot. 66% of MD has had 2+ shots. Another 7% has had only a single shot. Stats from usafacts.org.

I did use the 66% number, so I guess you mean the 7% 1-shot vs 10% that I rounded? If I rerun with that, I still get 83.3% vaccine effectiveness vs naive.

Lastly, these vaccination rates are from *now* not from back when the infections happened. Vaccination rate has risen over time.Each of these factors results in a lower amount of protection than you calculated.

If the breakthrough case rate being used is the recent one, then the vaccination rate over time isn't relevant to this calculation. If it's the total fraction over the entire pandemic, then yes this matters. I thought the 30% was the recent rate, not the total. I'll use https://coronavirus.maryland.gov/ as the source here for numbers.
This indicates that if we were to take all cases since the start of the pandemic then the percentage of deaths in breakthrough cases is 11.8%. If we take the numbers in August (the latest they report specific ratios for), 28% of the cases were breakthrough cases, so that's pretty close to the 30% I used. Granted I should use the vaccine counts from August in that case, which were 59% 2xVax and an additional 6% 1xVax at the start of August, and 61.4% 2xVax and an additional 6.7% at the end of August. In that case I get 78% effectiveness using the start-of-August numbers and 81% using the end-of-August numbers, so that does have a significant effect on the estimate. I'd run the September numbers but I'm not sure if they're available anywhere.

As for unreported cases, well...there is little we can do about data we don't have, it'd be pure speculation. Given that a *lot* of companies mandate weekly testing for all unvaccinated here, it seems highly probable that this number is fairly low. If any infections are being missed, it would more probably be among the vaccinated, who do not have this testing mandate. However, we have no good way to predict how large that effect is.

Infections since the beginning of the pandemic count for latent immunity, so frequent testing is only making those counts accurate if you were doing it since 2020. The 10% I'm using is only the reported cases, so it's a conservative estimate there. If we want to do more than wildly guess at the unreported cases, variations in test positivity fraction are the usual tool people try to use for that, but it's still going to give pretty soft estimates - it says more about relative fractions of missed cases than the absolute percentage.

As a footnote, we also have over 15% of the MD population with a booster. This hasn't been the case long enough to have good data on it yet, but if there is a significant infection risk variance between 2 and 3 shots, that might also be a complicating factor.

That will be a factor to consider going forward. That shouldn't be visible yet in the August numbers though, no?

Edit: Realized I didn't properly summarize. What this largely boils down to is a decent argument for waning protection against severe illness. Earlier time periods had remarkably lower death rates among the vaccinated, with one period being heavily advertised as wholly unvaccinated deaths. This appears to be significantly diminishing as time passes. While breakthrough infections were reasonably common even in the earlier timeframe, deaths were quite a bit rarer. We're sort of using death as a proxy for severe illness...but it's the outcome that is a fairly concrete metric and it's pretty objective, and preventing it is arguably the most important, or at least one of the most important purposes of a vaccine.

This probably means boosters will be tried. How well that works, well...that depends on a lot. Variants(some of which might not yet exist) are one potential wild card. We may also need to start redesigning vaccines. Either to account for variant specific changes, or to incorporate more viral features to make the vaccine less vulnerable to small variances. Natural resistance should be helping us along the way, though, so even if there's another spike, hopefully the overall intensity continues to recede.

Yes, there's certainly evidence for a degree of waning immunity, not arguing against that. Even if there weren't, it makes sense to try boosters. We use three-shot regimens for a lot of other diseases anyhow (MMR vaccine for example), and the third shot is what gets us from mid-90% effectiveness to the 99+% effectiveness needed to deal with the very high R0 of measles. The more layered protections we can manage, the better off we'll be.

But I guess it's a pet peeve of mine that people write news articles or caution pieces with superlatives and shock phrases like e.g. 'the vaccine provides excellent protection' or 'half of cases are breakthrough cases - the vaccine is a coin flip against delta' or whatever, and that feeds into this idea of wanting some one thing to be a silver bullet where you do that one thing and you can stop having to think about COVID ever again, and then when a useful tool ends up not being what they imagine they totally discount it. Not to mention that reporting 'percent effectiveness' is really misleading, because the thing that matters is actually 1 - that number. E.g. if you resist 100% of all damage, that's infinitely better than resisting 99% of all damage, not just 1% better. So I want to be precise about these analyses. It's more important to know as close as possible, quantitatively how much protection something gives so you can manage risk than to make a value judgment about whether the degree of protection should be praised or decried.

[Tyndmyr]

Yes, this is the point. If we're talking about the percentage of cases which are breakthrough, then that's the ratio of: (Cases in 2xVax)/(Cases in 2xVax + Other Cases)

So if there is some immunity in the Other Cases category, it will make the vaccine appear less effective than it actually is.

Well, of course. But we don't have a breakout for the 1x vax vs unvaccinated. They are all simply reported together. Putting any number on that is...just an assumption. Ends up being circular, if you assume the partial vaccination is effective, you get a result supporting overall effectiveness, if not...not. But both assumptions are just assumptions, not data.

I did use the 66% number, so I guess you mean the 7% 1-shot vs 10% that I rounded? If I rerun with that, I still get 83.3% vaccine effectiveness vs naive.

Correct, there's only 7% with just a single shot of the two shot vaccinations.

Single shot J&J is reported as fully vaccinated, as are those who got two shots.

If the breakthrough case rate being used is the recent one, then the vaccination rate over time isn't relevant to this calculation. If it's the total fraction over the entire pandemic, then yes this matters. I thought the 30% was the recent rate, not the total. I'll use https://coronavirus.maryland.gov/ as the source here for numbers.

The breakthrough rate is significant, but increasing. However, we are primarily tracking the breakthrough death rate. In June, this was 0%, as *all* the deaths were among unvaccinated.

The trendline for that is not great. I suppose one could argue that one or the other numbers are not perfectly accurate, and everything has a degree of error, but a sample pool of a state is pretty big. It seems highly unlikely that we would see such a trend by random chance.

This indicates that if we were to take all cases since the start of the pandemic then the percentage of deaths in breakthrough cases is 11.8%. If we take the numbers in August (the latest they report specific ratios for), 28% of the cases were breakthrough cases, so that's pretty close to the 30% I used. Granted I should use the vaccine counts from August in that case, which were 59% 2xVax and an additional 6% 1xVax at the start of August, and 61.4% 2xVax and an additional 6.7% at the end of August. In that case I get 78% effectiveness using the start-of-August numbers and 81% using the end-of-August numbers, so that does have a significant effect on the estimate. I'd run the September numbers but I'm not sure if they're available anywhere.

We don't have the same precision. However, the 40% number I cited originally comes from one of Hogan's advisors who made a statement on Fox News. https://townhall.com/tipsheet/katiep...eaths-n2597628

That is probably an approximation, but it's a late Oct source, so the preceding 6-8 week timeframe would include Sept. If accurate, this indicates a continuation of the increase we're seeing in the deaths from breakthrough cases.

Infections since the beginning of the pandemic count for latent immunity, so frequent testing is only making those counts accurate if you were doing it since 2020. The 10% I'm using is only the reported cases, so it's a conservative estimate there. If we want to do more than wildly guess at the unreported cases, variations in test positivity fraction are the usual tool people try to use for that, but it's still going to give pretty soft estimates - it says more about relative fractions of missed cases than the absolute percentage.

Yeah, there's an effect here, but we can't truly measure it. Testing has generally increased over this time period, and we probably cannot know for sure the impact more frequent testing has on certain subpopulations since we're working with...imprecise data on that. I can say with certainty that for at least quite a few months, the unvaccinated have been tested *more* frequently, but teasing out exactly how strongly that affects missed cases is probably not possible with the data we have.

That will be a factor to consider going forward. That shouldn't be visible yet in the August numbers though, no?

Unclear. There is a substantial population here that opted for vaccine tourism, hopping across state lines to get boosters before they were authorized. This is difficult to track because they were deliberately avoiding that, but I believe I linked a source a ways upthread indicating that it was significant enough for the CDC to take into account for safety purposes.

It's a possible source of error, but difficult to quantify.

Yes, there's certainly evidence for a degree of waning immunity, not arguing against that. Even if there weren't, it makes sense to try boosters. We use three-shot regimens for a lot of other diseases anyhow (MMR vaccine for example), and the third shot is what gets us from mid-90% effectiveness to the 99+% effectiveness needed to deal with the very high R0 of measles. The more layered protections we can manage, the better off we'll be.

Eh, if efficacy were quite high and waning immunity weren't a thing, two shots would be fine. But that seems unlikely to be the case.

The fly in the ointment is...how much of the waning immunity comes from decreased antibody effectiveness over time, and how much comes from variants?

Because the former you can kind of attack with boosters....but that may not end at three boosters. The Tetanus shot, for instance, generally wanes enough that it requires another shot every ten years-ish. They don't usually bother topping it up unless you've had possible exposure, but if you have, well, you'll likely need a shot if yours is old enough. The rate of decay for covid appears to be a *lot* faster. So, booster tempo needed to address that is...high. This imposes some logistical problems that one time vaccinations do not. It also is really rough in terms of long term efficacy for cases like the transplants. Getting a shot beforehand may protect for a time, but if that protection fades, you're back in that position of choosing in a while.

The latter is a problem because well, boosters may not fix it. There might be some overlap, but a variant that is poorly targeted by the vaccine may not see amazing improvement from more of the same. At that point, new development is needed, akin to what we do for flu shots.

[NichG]

Eh, if efficacy were quite high and waning immunity weren't a thing, two shots would be fine. But that seems unlikely to be the case.

Depends on what your risk acceptance is, the prevalence of Covid, etc...

Lets say one wants to maintain a 1% per year risk of getting Covid, and we end up endemic around the current numbers in Maryland. Then...

- With a 50% effective vaccine, you could spend about 10 hours per week indoors with other people no closer than 2 meters being silent or talking quietly if everyone is wearing cloth masks.

- With a 85% effective vaccine, you could work 8 hours a day in those conditions while maintaining my risk acceptance. Or you could spend about 1.5 hours a week in a bar talking with other people unmasked and eating/drinking, but not also be able to include 8 hours a day working in person.

- To work in those bar conditions for 40 hours a week and maintain that risk budget, you'd need a 99.5% effective vaccine (or wear a tight-fitted N95 mask and have a 95% effective vaccine).

[Tyndmyr]

Depends on what your risk acceptance is, the prevalence of Covid, etc...

Sure. That's going to vary a lot by person, I imagine. Age and other factors seem like they'd greatly affect risk tolerance.

Lets say one wants to maintain a 1% per year risk of getting Covid, and we end up endemic around the current numbers in Maryland. Then...

- With a 50% effective vaccine, you could spend about 10 hours per week indoors with other people no closer than 2 meters being silent or talking quietly if everyone is wearing cloth masks.

Tolerance for this is gone. Most counties no longer have indoor mandates, and those that do are repealing them...and in any case noncooperation has become increasingly common, despite masking being nigh-universal earlier.

- With a 85% effective vaccine, you could work 8 hours a day in those conditions while maintaining my risk acceptance. Or you could spend about 1.5 hours a week in a bar talking with other people unmasked and eating/drinking, but not also be able to include 8 hours a day working in person.

- To work in those bar conditions for 40 hours a week and maintain that risk budget, you'd need a 99.5% effective vaccine (or wear a tight-fitted N95 mask and have a 95% effective vaccine).

In practice, people sort of have to work. Some jobs are work from home, but most are not, and mine certainly isn't. I don't think most folks have the money to just not work for an arbitrary length of time. I suspect the primary outcome is just "risk tolerance lowers."

I suspect that any measure can only be maintained for so long before it outstays its welcome, and the more intrusive the measure, the shorter that time generally is. Lockdowns have largely ended, and remaining policies from that time are fading. Yeah, a tight fitted N95 is, by the numbers, the safe bet. I don't see many of those around, though. Masks are getting rarer, and where still worn, are chosen largely for comfort, not efficacy.

So, where's that lead? We can *try* a lot of boosters, but unfortunately, mandates have badly hurt the tolerance for that. This area was strongly pro-vaccine, but this has become...contentious. I think booster shots every six months would be a difficult option to persuade people to pursue. Other measures are also, at this point, a hard sell.

I am a bit concerned that we're approaching a point where there is fairly little further that can be done to mitigate...well, anything, really. Perhaps pandemics ought to be approached differently somehow...and here I am of course speculating. Would a slower, but more comprehensive vaccine have been a better bet overall? You lose on the front end, but maybe gain on the back. Maybe. Variants are inherently challenging to predict.

Earlier detection and slowing of spread is of course desirable, but how to achieve that is something of a challenge. There's always a new bug every so often, and most of them do stop before they hit this level. Hindsight means we're looking at the case where that failed, and there is always a possibility of containment failing.

I do hope the tradition of more elbow room in lines and stuff lasts, at least. I kind of like the personal space bubble, and that seems generally useful against illness of any sort.