Also note efficacy can be misleading. Pfizer was 100% effective at preventing death or serious illness in their trials, which is what's really important. It was "only" 95% effective against getting the disease at all, which isn't particularly important.

i.e. even a vaccine that has 0% efficacy (doesn't prevent mild cases at all) and yet is ~100% at preventing severe cases would be on par with Pfizer's, whereas one that does the converse (prevents all mild cases but no severe ones) would be drastically worse. Not saying these extremes are likely, but we need numbers on the former rather than the latter to say whether another vaccine truly 'rivals' that of Pfizer.

P.S. (rant): Why don't we just define a single word for the metric for what really matters and just use that?? It boggles my mind.

Another example: Sinovac's Vaccine is 100% effective (at preventing death in Brazil) and 50.7% effective (at preventing symptoms in Brazil)

Coronavac efficacy in Brazil has been way higher than 90% at preventing death, right: https://www.bbc.com/portuguese/geral-55639265 ?

Yep, fixed. Copied the wrong number.

I believe it is 95% effective at getting the disease with symptoms. I understand they never tested people who had no symptoms in their trials.

Oh interesting. Would you have a link on this?

That's what I have seen mentioned repeatedly on HN. Some of the comments link to articles and studies, but not being a doctor myself, I have not verified them

https://hn.algolia.com/?dateRange=pastYear&page=0&prefix=fal...

Ok thanks. Btw, I think my comment is actually correct with that definition, given how having the disease (COVID-19) implies having symptoms. IIRC that's precisely when they say you have SARS-CoV-2 (the virus, which you'd be an asymptomatic carrier of), but not COVID-19 (the disease).

The main problem is that you need a lot more data to make confident statements about severe cases than just about symptomatic cases. That 100% number of the Pfizer trial had a huge uncertainty interval. So even though severe cases are what you care most about, you usually don’t want to wait for a study that gives you good data on those and measure (and compare) symptomatic cases instead. Zeynep Tufekci discussed this issue of study endpoints in one of her recent articles: https://www.theinsight.org/p/vaccine-efficacy-statistical-po...

But couldn't they just tell people about the uncertainty? (like "90-98% effective"?) Or underestimate the efficacy (in some objective way, like "at least 90% effective"). Or something else, as long as they're ultimately measuring what matters and what people actually care about. It's not like that 100% with their sample sizes would realistically ever suddenly turn into 50%, would it? It seems just inherently wrong to measure the wrong thing and convey it with certainty than to measure the right thing and convey it with the associated uncertainty, does it not?

They actually had a table on the efficacy for severe cases in the supplemental material of the original Pfizer study (table S5 in [1]), but all it shows is that the study doesn’t really give any information on this. The 95% confidence interval given for severe cases after the first vaccine is 20.1–99.7%.

From what I understand, it’s really a trade-off between waiting for proper data about the efficacy you’re most interested in, or starting to vaccinate people as soon as you have efficacy data for an endpoint that’s still meaningful but easier to reach. Having much fewer symptomatic cases is already a pretty good result, and I think it was expected that the protection would also extend to severe cases.

[1]: https://www.nejm.org/doi/suppl/10.1056/NEJMoa2034577/suppl_f...

Preventing infection is very important in reducing spread of the disease to others. Vaccines in the end work best when we reach herd immunity, and that occurs by preventing most infection, and thus the possibility of mutation.

> Preventing infection

I agree! I didn't say anything about the importance of preventing infection.

Getting the disease is being infected. It is important that people not get it in the first place, so they can't transmit it to others.

Not quite. As far as the terminology goes, I don't think they're the same, since to me having disease implies feeling ill—note "disease" = "not at ease" = "COVID-19"—whereas infection implies merely carrying the SARS-CoV-2 virus in your body regardless of any visible effects. (Having SARS-CoV-2 does not imply you have COVID-19!) But in any case, let me clarify this.

What I agree with (and in my initial comment I said nothing about) is that it matters if people can spread the virus to others.

What I don't with is that it matters if you're "infected" if you can't spread it to others (and don't have severe symptoms).

The important thing here is that, from what I recall Dr. Fauci said a month or so ago, there have been promising indications that vaccinated people who get infected with the virus are both significantly less likely to get the disease (COVID-19) and to transmit the virus (SARS-CoV-2) to others (independently of the disease, COVID-19). You can read about this on the CDC website [1]. (And my assumption was that was a big factor in why they lifted the mask mandate... because fewer people are capable of spreading it even if they get the virus.)

So no, it doesn't matter if you merely get infected. What really matters is how many people can become severely sick as a result of your infection, whether that's you or others. Which is important because these vaccines seem to limit transmission itself.

[1] https://www.cdc.gov/coronavirus/2019-ncov/science/science-br...

If you have the virus in you, there is always potential for illness, transmission, and mutation. Even if chances are reduced, not getting it in the first place is better; and all else equal, a vaccine that does a better job of stopping infection is better than one that does not. Since virus mutations are an unknown, "past performance does not guarantee future results" very much applies. So although I agree how sick people get is very important for a vaccine, and in the short term certainly the most important aspect, it is not the only important criterion to measure.

As to why the US CDC lifted the mask mandate while the WHO is recommending vaccinated people wear masks to try to stop the spread of the Delta variant, and there have been many documented cases of large parts of groups of vaccinated people getting the Delta variant as well as other variants - some of whom have gotten seriously ill - I have no comment.

You're splitting hairs at this point. Yes, all else being equal, I'd prefer not to get the virus too. All else being equal there are lots of things I'd prefer over other things, but they turn out to be negligible and stop mattering in the presence of so many bigger fish to fry.

And I'm not saying they should stop measuring anything either. They should measure whatever they feel like. All I'm saying is most people's #1 concern (including that of those in healthcare) is "will I/others get severely ill", and IMHO that means that's the information that needs to be conveyed as accurately as possible to the public first and foremost. Especially when there seems to be such a highly likely possibility that many people who merely "have" the virus will feel no effects. Of course nothing is a guarantee and it's even possible the vaccines will turn out even worse than the disease in a few years. Nobody claimed otherwise.

Well this has gotten old.

You stated that the current form of measurement was not very good. I pointed out that we care about transmission and mutation for long term eradication of the disease - hence the measurement that is used. It's obviously not negligible or the entire professional world would not use it as a measuring stick for vaccine effectiveness.

My comment wasn't about what measures the profession uses to do their work. My comment was about what measure they're communicating to us, random members of the public. These need not be the same thing.

That is fair enough. However it would still backfire, because the general public can't listen to anything responsibly.

If we report the vaccine is 99.5% effective at preventing serious illness, people will not listen to ANYTHING about being cautious and wearing a mask after they are vaccinated. I predict a nasty wave of COVID this fall/winter including in relatively highly vaccinated areas due to just this attitude already.

> If we report the vaccine is 99.5% effective at preventing serious illness, people will not listen to ANYTHING about being cautious and wearing a mask after they are vaccinated.

Uhm, the CDC literally told vaccinated people they don't need masks indoors or outdoors. [1] At that point it's kind of silly to say quoting 99.5% does or will a detrimental impact.

> I predict a nasty wave of COVID this fall/winter including in relatively highly vaccinated areas due to just this attitude already.

No; if it happens, it certainly wouldn't be due to this; at best ('best' for your argument), it would be due to their overriding guidance re: not wearing masks, which I mentioned above. However, I predict if this happens, it'll be due to changing circumstances like worse variants breaking through the vaccine (which the CDC admits they do not have sufficient data on [2]), not due to their current advice being bad due for current variants.

[1] https://www.webmd.com/lung/news/20210513/cdc-vaccinated-you-...

[2] https://www.cdc.gov/coronavirus/2019-ncov/vaccines/fully-vac...

How much data (read: jabs) is necessary to evaluate the ability to prevent hospitalizations when even without the jab only a small percentage of cases lead to hospitalization? How much care is given to ensure the jab group's demographics reflect the "control" (i.e., the broader population).

For example, ppl getting chemo aren't given the jab, are they? Yet they are counted if they get the virus, end up in hospital, etc. The point being, extrene Covid preys on the (immune) weak. Yet the jabs aren't tested on ppl who profile as such. How is that accounted for in evaluating a vaccine?

Disclaimer: layman here, just repeating what I heard and read the past months/years.

> How much care is given to ensure the jab group's demographics reflect the "control" (i.e., the broader population).

This is mostly done through sampling all ages. In general, older people have a weaker immune system than younger, so older people can serve as a model for younger, immune-suppressed people. That's not ideal, but generally preferable to testing new drugs in chemo patients.

In cases where people have an elevated risk (chemo, pregnancy, other immune suppression situations like after transplants), the best you can do is vaccinate as many of the contact persons around them.

(Regarding pregnancies, I don't know how much of that is actual risk regarding vaccination, and how much is related to the difficulty of testing drugs on pregnant women. For example in Germany, drug tests are on pregnant women are a very delicate topic and mostly avoided by pharma companies, mostly due to https://en.wikipedia.org/wiki/Thalidomide_scandal )

I'm a layman so I don't know if I can really give a satisfying answer to your question. All I know is what I see just doesn't make sense to me.

Even in April 2020 we had rough but useful numbers on the hospitalization rate. Just some quick Googling turns up e.g. [1] which says that around 20% of those who test positive end up hospitalized. Now the Pfizer vaccine by itself had like 40k participants, with 9 severe (out of 162 confirmed cases) of COVID-19 in the placebo group, and 1 severe case (out of 8 confirmed cases) in the vaccine group. [2] [3] So right off the bat it looked pretty promising, and IMHO they could've just cited these numbers and let people judge for themselves... it's not like people don't have the intuition to realize that 1 patient couldn't have been 2-3 instead. But even if you think that wasn't enough data to be convincing, at least now we have mass vaccination rates they could be citing. Like literally ~40k people have been hospitalized with COVID-19 in just the last week and nearly all COVID deaths are among the unvaccinated now, and it's not like the vaccine just came out last week.

An Associated Press analysis of available government data from May shows that “breakthrough” infections in fully vaccinated people accounted for fewer than 1,200 of more than 853,000 COVID-19 hospitalizations. That’s about 0.1%. [4]

How is that still not enough data to switch metrics at this point? Why do they still think efficacy rates are more relevant for people instead?

[1] https://www.tfah.org/wp-content/uploads/2020/04/COVIDunderly...

[2] https://www.pfizer.com/news/press-release/press-release-deta...

[3] I don't know why this reporting differs from the 100% I see claimed elsewhere; 1 severe case doesn't sound like 100% effective at preventing severe cases? I don't know if they're discussing different trial phases or have different thresholds for "severe" or something else... but regardless, it's irrelevant to my point.

[4] https://apnews.com/article/coronavirus-pandemic-health-941fc...

I think the 100% number comes from the fact that the one vaccinated severe case in the original study just was just barely below the threshold of oxygen saturation they had defined but the person required no medical care, so it wouldn’t be classified as severe by “regular people’s standards”.

Ah, thanks!

More questions, if you don't mind...

Is it really scientific (?) to compare the first 6 months to the second? That is, for example, Covid ravaged extended care facilities. Was that the virus of too many facilities being ill prepared?

As for recently, "unvaccinated" is fairly broad and in aggregate hides important details. That is, are those ppl unjabbed by choice, or by necessity (i.e., immune compromised, pregnant, nursing homes, etc.) Why isn't this important distinction made?

I'm not a CDC spokesman, sorry. You'll have to ask them. All I can do is to try to exercise a bit of common sense. And my common sense says you shouldn't wait for an RCT for "science" to tell people parachutes save lives...

Marketing

Absolutely, this number may be confirmed when tested by a third party. Medias were cautious on the russian's vaccine claims until the Lancet published a study on its efficacy.

Edit: grammar

The Lancet published Andrew Wakefield's study on the connection between vaccines and autism, and the recent study on hydroxychloroquine based on entirely fabricated data.

The Lancet publishing data does not mean the data is correct, they do not verify it. It's all based on trust.

I think you have a good point. Not sure why you're downvoted or "almost invisible".

I agree too. I look forward to Cuba allowing third parties to corroborate their results.