Phizer and Moderna are going to come to the rescue of a select few. Let them do so at the best level available. The proof of concept they have shown will likely lead to a cascade of positive results, of varying CI%. We will need them all, and even if the general roll-out is late, choice will come into play when people are reluctant to take a vaccine that 1) can give a strong allergic reaction 2) was derived from stem-cell lines 3) was made from GMO's 4) is from a company they don't like 5) you get the idea. Whatever it takes, we need to effectively vaccinate as many people as possible.
Personally, I think the avoidance numbers will be much lower than the polls are predicting, because of a feature of human nature: When people see other sub-groups getting in line ahead of them, their jealousy will incline them to want to join the herd.
To accept a smaller dose of the vaccine, wouldn’t it also be important to know more about what that does to transmission rates? It seems to me people who get a vaccine will surely (?) ease up on their precautions, so it might be even more risky to ease up on the strength of the vaccine without knowing this.
It's a challenge for communicating for sure! Plus, some immunologists (including chief medical officer of Moderna) are suggesting that the first dose could act as the prime.
Zeynep, in the unlikely case you aren't already doing this, can you please try to get this into the mainstream press? What I'm seeing so far in the media -- particularly here in Germany -- is dominated so heavily by "do it right or bust"-type thinking and a fascination for ethical conundra over practical solutions, I'm worried that we won't get the vaccine until Fall. It takes a certain kind of person to believe that rolling out the vaccine needs to wait until panels of ethicists have finished deciding who gets it first (the elderly or medical workers?), but somehow this kind of person tends to be particularly strongly represented in newsrooms and (to a lesser extent) in the German government. Someone needs to state the obvious: something is better than nothing, and now is better than later.
I think that Pfizer and others don't want the responsibility of choosing between 1 and 2 doses and dealing with lower effectiveness; so is this something that policymakers want to make a decision upon? Part of the no testing strategy is to avoid having the data condemn you; why would they change that now?
I get their motivations. But given mass vaccinations are about to start, we could push for this? (The companies are also getting a lot of flexibility and exceptions).
I find it surprising that the contagion part isn't better explained to the public. We don't know how much, but most everyone in the field expects a big impact on transmission. Of course, good to wait for data before relaxing.
A somewhat differing view: In the latest iteration (#134) of his podcast ( https://www.mdr.de/nachrichten/podcast/kekule-corona/kekule-corona-zweite-impfung-viruserkrankungen-vegane-ernaehrung-audio-100.html ), the German virologist Alexander Kekulé has commented on the one-dose plan. The TL;DR is that he is somewhat worried but is not dismissing it as a bad idea. He believes the possibility of "escape mutation" via half-immunity is real, and the use of existing studies for one-dose effectivity is somewhat hampered by low statistical power. (At least this is how I understand him; there is no transcript yet that I could link to.) Note that he is talking not just about the mRNA vaccines but also about the Oxford/AZ one (which he doesn't fully trust yet). His own upshot is to give the elderly the original two-dose program and everyone else the "first dose now, second dose when it's ready" treatment. (He argues that it's particularly the elderly and immunocompromised that are likely to contribute to escape mutations.)
Kekulé's podcast has been a great source of information and informed opinion throughout the last year, so I am inclined to take him far more seriously on this than anything written by a philosopher or politician. As they say on reddit, Δ (but not to the opposite).
We will know soon enough, but the chief medical officer of Moderna hinted at all this at an investor call. In fact, I'm really hoping we get better data from Moderna on this, which would really strengthen the case for an immediate trial.
Zeynep asks what's better, one dose of vaccine for 200 million people with 60-80% efficacy or two doses of vaccine for 100 million people with 95% efficacy? If 70% of unvaccinated people will ultimately get COVID-19, the standard two-dose regimen would lead to 5 million vaccinated people and 70 million unvaccinated people becoming infected and the suggested one-dose regimen would lead to 40-80 million people becoming infected. The other factors she mentions, including public trust, must be used to decide how best to use scarce vaccine.
One other consideration is the severe disease prevention. If one dose reduces to a much less risky disease, it would be yet another consideration (beyond any disease which is what the top-line efficacy numbers are reporting). Strongly agree about the public trust part.
Good point about risk reduction, especially if the AstraZeneca vaccine program continues to have problems. What if the company cannot supply the 300 million doses that the US has contracted to purchase? Then shifting to a one-dose regimen might be the most prudent course.
This is certainly a very good line of thought to have now, and to try to obtain more data now as well. One aspect that is missing in your discussion, though, or which I didn't see :), is that lowering the infection rate has not to be the ultimate goal, but most important is to save as many lives as possible. So I think for the most vulnerable, who should be highest priority, we should always play safe and provide them with two shots, and probably also for the people directly surrounding them. Then, and only then, it makes sense to turn the attention towards the infection rate. Since this will probably only happen some month from now (at least this is what I understand will be happening here in Europe) there should be enough time to collect more data as you suggest.
Okay, I read the paper. >18,000 people took the second shot of vaccine, and >18,000 took the second shot of saline, so that is a relatively robust N showing that the first shot was 52% effective.
Conversely, only 304 of the vaccine recipient group failed to get the second shot of vaccine, while 316 people failed to get the second shot of saline. I think this is the group that got the 82%.
I like to find out when I am wrong, and what the right answer is, so if anyone wants to straighten me out, don't worry about my feelings!
Maybe the N that was too small was the N for the "one shot only" treatment. I would be more inclined to have confidence in the CI number derived from the highest # of recipients. But seeing that the "in between shots" subset only scored 52%, maybe that's closer to where one shot only actually will get you. I don't like doubting data (not smart!), but those two figures are not really compatible.
I'm confused by the difference between the "After dose 1" and "After dose 1 to before dose 2" groups. Does the former consist of participants who bailed out after getting their first dose? Or do the two groups just represent two parts of the 21-day interval between dose 1 and dose 2? If so, where is the breaking point?
(This all still looks more like bad writing than any kind of misconduct to me, and I still am far from convinced that the vaccine is useless without a booster.)
I think the data is difficult to understand for one dose treatments because Phizer/Biontech designed their study as a two-dose study. To really check one dose efficacy, they would have to do a large study of just one dose.. In this study, they only have incidental data for one dose, and not that much of it. At least it looks that way to me, I'm not an epidemiologist, or anything close. They may have done so at the stage one level, decided it didn't meet their standard.
Still confused, though. According to this flowchart, only 316 participants received placebo dose 1 but no dose 2. According to Figure 3 (embedded in Zeynep's post), 275 of them got sick. 275 out of 316, seriously? Was there a super-spreader event somewhere near that study?
If someone got COVID between dose 1 and dose 2 and was too sick to get the second dose, which group would he be counted into? What does "Had ongoing or pending status" mean? What about "Were no longer eligible"?
Alright. Sorry for the message ziggurat, but I have a hunch what's going on now. (To make it clear: a hunch, not an answer! But a good enough hunch to make me stop worrying about the single dose.)
There are 21 days between doses 1 and 2.
- Imagine you got sick *shortly* after dose 1. Then, when the 21 days are over, you're probably well enough to get dose 2, and you end up in the "After dose 1 to before dose 2" group.
- Now imagine you got sick closer to the end of the 21 days. Then, when it's time for dose 2, you're too sick to get it, and thus withdraw from the study. So you end up in the "After dose 1" group.
And now it all makes sense: The vaccine takes a while to work (immune responses require a few weeks), so its efficacy in week 1 is not great. Those participants end up in the "After dose 1 to before dose 2" group. Its efficacy in weeks 2-3 is better, and those end up in the "After dose 1" group.
Hold the phone, what's that second result line about? Result CI from between first shot and second shot is only 52%, with a wicked-big range of 29-64. I can't think of a good reason that this would vary so widely from the one treatment only study. In fact, they should be pretty close, with the after first-shot, before second shot being more likely to be HIGHER CI% than one shot only, because less elapsed time to get SARS CoV-2 before booster. Is this just because data are not robust? 21,000 is a pretty good number, but what was the N for just the "in between shots" treatment?
Heh. Honestly, the confidence interval is huge but I think the bigger question is durability. The immunologists I spoke with were a lot more confident about the first month than what happens in the longer run.
I'll have to take their word for it. You showed a lot of guts to run that opinion in the NYT, doesn't seem to have done any harm. Looks like the 2 dose decision was made very early on, so that's what we are stuck with, research wise. J&J is next in line, and if they get approved (when), that will really accelerate the number of people getting doses. What bothers me is that they are holding back half the Phizer/Moderna doses for the first tranche, to assure a second dose. This is where I would like to see a more adventurous use of resources to 'double' the # of recipients. So much worry about exact timing of 2nd dose! Why?
On this topic, I'm planning on listening to what the epidemiologists suggest. At the moment, it doesn't seem like there is sufficient peer reviewed opinion to suggest that the single dose route is appropriate.
She's an immunologist (even more appropriate!) but it doesn't answer the question at hand—the case for the booster is clear, but the trade-off isn't a pure immunology question because of the shortage. At some point, we will have data on single dose and if it is indeed reasonable and efficacious for at least a few months, we will be able to go back and calculate excess deaths. Hence the thorniness.
Another challenge would to be how a single-dose option might be received in under-served (and/or historically badly-served) communities. Does the single-dose option look like an abusive cost-cutting strategy, wrapped up in some marketing re: more lives saved?* Some folks get the low cost option w/higher risk; others get the high quality/cost full treatment w/lower risk? (Similar to the problem of Trump-adjacent people getting treatments not generally available.)
There are some communities who are vaccine hesitant due to historical medical missteps and worse. There are ethical issues re: treating them fairly and forthrightly. I'd want to get input from key players in those communities re: the single-dose idea. Regaining their trust need to be part of the equation.
* I know this is not where you are coming from. It is however, a possible interpretation for someone who lacks trust in the medical system.
Yes, it's a thorny case of how to reduce overall harm.
I'm open to the idea, but after all of the unfortunate Covid-related political decisions (hydroxychloroquine, etc), I'm firmly in the camp of go slow and behave conservatively. Esp when in a situation w/so many unknowns.
From The New York Times:
Can We Do Twice as Many Vaccinations as We Thought?
By Zeynep Tufekci and Michael Mina
Data suggests significant protection even without a second shot. If studies prove that’s true, it could be a game changer.
https://www.nytimes.com/2020/12/18/opinion/coronavirus-vaccine-doses.html?smid=em-share
Yep! Trying to push the idea. It's been a complicated year; some amazing science but also such slow response at times to the emerging science.
Another thought: there are going to be many, many more vaccines available, and some of those will be easier to make, cheaper to make, easier to store, and cheaper to ramp-up. https://www.nytimes.com/interactive/2020/science/coronavirus-vaccine-tracker.html
Phizer and Moderna are going to come to the rescue of a select few. Let them do so at the best level available. The proof of concept they have shown will likely lead to a cascade of positive results, of varying CI%. We will need them all, and even if the general roll-out is late, choice will come into play when people are reluctant to take a vaccine that 1) can give a strong allergic reaction 2) was derived from stem-cell lines 3) was made from GMO's 4) is from a company they don't like 5) you get the idea. Whatever it takes, we need to effectively vaccinate as many people as possible.
Personally, I think the avoidance numbers will be much lower than the polls are predicting, because of a feature of human nature: When people see other sub-groups getting in line ahead of them, their jealousy will incline them to want to join the herd.
I agree that I think the take-up will be excellent, to be honest.
To accept a smaller dose of the vaccine, wouldn’t it also be important to know more about what that does to transmission rates? It seems to me people who get a vaccine will surely (?) ease up on their precautions, so it might be even more risky to ease up on the strength of the vaccine without knowing this.
It's a challenge for communicating for sure! Plus, some immunologists (including chief medical officer of Moderna) are suggesting that the first dose could act as the prime.
Zeynep, in the unlikely case you aren't already doing this, can you please try to get this into the mainstream press? What I'm seeing so far in the media -- particularly here in Germany -- is dominated so heavily by "do it right or bust"-type thinking and a fascination for ethical conundra over practical solutions, I'm worried that we won't get the vaccine until Fall. It takes a certain kind of person to believe that rolling out the vaccine needs to wait until panels of ethicists have finished deciding who gets it first (the elderly or medical workers?), but somehow this kind of person tends to be particularly strongly represented in newsrooms and (to a lesser extent) in the German government. Someone needs to state the obvious: something is better than nothing, and now is better than later.
I was worried myself about missing my booster shot (due to job-related travel) -- I feared it could somehow help the virus adapt, in a "what doesn't kill me makes me stronger" way, like stopping antibiotics courses supposedly does to bacteria. Some google searches later, not only doesn't this seem to be a thing, but apparently even the oft-repeated claim about antibiotics isn't actually true ( https://www.who.int/news-room/q-a-detail/antimicrobial-resistance-does-stopping-a-course-of-antibiotics-early-lead-to-antibiotic-resistance , https://jamanetwork.com/journals/jamainternalmedicine/article-abstract/2536180 ). It seems that, as far as medical topics are concerned, scary what-ifs spread much faster than facts.
I think that Pfizer and others don't want the responsibility of choosing between 1 and 2 doses and dealing with lower effectiveness; so is this something that policymakers want to make a decision upon? Part of the no testing strategy is to avoid having the data condemn you; why would they change that now?
I get their motivations. But given mass vaccinations are about to start, we could push for this? (The companies are also getting a lot of flexibility and exceptions).
Yes, this is probably the best path; the challenges of getting good information in terms of fatality are exposed in this thread by Zachary Lipton: https://twitter.com/zacharylipton/status/1337121325492604932 (paper here: https://arxiv.org/abs/2012.04825 ). And we still don't know the impact in contagion!
So the lack of testing will also impact decisions on this level as well.
I find it surprising that the contagion part isn't better explained to the public. We don't know how much, but most everyone in the field expects a big impact on transmission. Of course, good to wait for data before relaxing.
A somewhat differing view: In the latest iteration (#134) of his podcast ( https://www.mdr.de/nachrichten/podcast/kekule-corona/kekule-corona-zweite-impfung-viruserkrankungen-vegane-ernaehrung-audio-100.html ), the German virologist Alexander Kekulé has commented on the one-dose plan. The TL;DR is that he is somewhat worried but is not dismissing it as a bad idea. He believes the possibility of "escape mutation" via half-immunity is real, and the use of existing studies for one-dose effectivity is somewhat hampered by low statistical power. (At least this is how I understand him; there is no transcript yet that I could link to.) Note that he is talking not just about the mRNA vaccines but also about the Oxford/AZ one (which he doesn't fully trust yet). His own upshot is to give the elderly the original two-dose program and everyone else the "first dose now, second dose when it's ready" treatment. (He argues that it's particularly the elderly and immunocompromised that are likely to contribute to escape mutations.)
Kekulé's podcast has been a great source of information and informed opinion throughout the last year, so I am inclined to take him far more seriously on this than anything written by a philosopher or politician. As they say on reddit, Δ (but not to the opposite).
Do we know if the Moderna vaccine has the same properties in terms of effectiveness after 1 dose?
We will know soon enough, but the chief medical officer of Moderna hinted at all this at an investor call. In fact, I'm really hoping we get better data from Moderna on this, which would really strengthen the case for an immediate trial.
Zeynep asks what's better, one dose of vaccine for 200 million people with 60-80% efficacy or two doses of vaccine for 100 million people with 95% efficacy? If 70% of unvaccinated people will ultimately get COVID-19, the standard two-dose regimen would lead to 5 million vaccinated people and 70 million unvaccinated people becoming infected and the suggested one-dose regimen would lead to 40-80 million people becoming infected. The other factors she mentions, including public trust, must be used to decide how best to use scarce vaccine.
One other consideration is the severe disease prevention. If one dose reduces to a much less risky disease, it would be yet another consideration (beyond any disease which is what the top-line efficacy numbers are reporting). Strongly agree about the public trust part.
Good point about risk reduction, especially if the AstraZeneca vaccine program continues to have problems. What if the company cannot supply the 300 million doses that the US has contracted to purchase? Then shifting to a one-dose regimen might be the most prudent course.
This is certainly a very good line of thought to have now, and to try to obtain more data now as well. One aspect that is missing in your discussion, though, or which I didn't see :), is that lowering the infection rate has not to be the ultimate goal, but most important is to save as many lives as possible. So I think for the most vulnerable, who should be highest priority, we should always play safe and provide them with two shots, and probably also for the people directly surrounding them. Then, and only then, it makes sense to turn the attention towards the infection rate. Since this will probably only happen some month from now (at least this is what I understand will be happening here in Europe) there should be enough time to collect more data as you suggest.
I think there's data and reasons you outline that a trial should not include high-risk people.
Okay, I read the paper. >18,000 people took the second shot of vaccine, and >18,000 took the second shot of saline, so that is a relatively robust N showing that the first shot was 52% effective.
Conversely, only 304 of the vaccine recipient group failed to get the second shot of vaccine, while 316 people failed to get the second shot of saline. I think this is the group that got the 82%.
I like to find out when I am wrong, and what the right answer is, so if anyone wants to straighten me out, don't worry about my feelings!
Maybe the N that was too small was the N for the "one shot only" treatment. I would be more inclined to have confidence in the CI number derived from the highest # of recipients. But seeing that the "in between shots" subset only scored 52%, maybe that's closer to where one shot only actually will get you. I don't like doubting data (not smart!), but those two figures are not really compatible.
I'm confused by the difference between the "After dose 1" and "After dose 1 to before dose 2" groups. Does the former consist of participants who bailed out after getting their first dose? Or do the two groups just represent two parts of the 21-day interval between dose 1 and dose 2? If so, where is the breaking point?
(This all still looks more like bad writing than any kind of misconduct to me, and I still am far from convinced that the vaccine is useless without a booster.)
I think the data is difficult to understand for one dose treatments because Phizer/Biontech designed their study as a two-dose study. To really check one dose efficacy, they would have to do a large study of just one dose.. In this study, they only have incidental data for one dose, and not that much of it. At least it looks that way to me, I'm not an epidemiologist, or anything close. They may have done so at the stage one level, decided it didn't meet their standard.
Ah, here are the explanations for why people did not get dose 2:
https://www.nejm.org/na101/home/literatum/publisher/mms/journals/content/nejm/0/nejm.ahead-of-print/nejmoa2034577/20201210-05/images/img_medium/nejmoa2034577_f1.jpeg
Still confused, though. According to this flowchart, only 316 participants received placebo dose 1 but no dose 2. According to Figure 3 (embedded in Zeynep's post), 275 of them got sick. 275 out of 316, seriously? Was there a super-spreader event somewhere near that study?
The 275 sick in the placebo group includes people who got a second placebo dose.
Why would they be in the "After dose 1" group then (as opposed to "After dose 1 to before dose 2")?
“After dose 1 to before dose 2” is a subset of “after dose 1”. The chart lists all participants in the study who got COVID.
Notice that 82 + 21 + 172 = 275.
Aaaah! That explains things.
If someone got COVID between dose 1 and dose 2 and was too sick to get the second dose, which group would he be counted into? What does "Had ongoing or pending status" mean? What about "Were no longer eligible"?
Alright. Sorry for the message ziggurat, but I have a hunch what's going on now. (To make it clear: a hunch, not an answer! But a good enough hunch to make me stop worrying about the single dose.)
There are 21 days between doses 1 and 2.
- Imagine you got sick *shortly* after dose 1. Then, when the 21 days are over, you're probably well enough to get dose 2, and you end up in the "After dose 1 to before dose 2" group.
- Now imagine you got sick closer to the end of the 21 days. Then, when it's time for dose 2, you're too sick to get it, and thus withdraw from the study. So you end up in the "After dose 1" group.
And now it all makes sense: The vaccine takes a while to work (immune responses require a few weeks), so its efficacy in week 1 is not great. Those participants end up in the "After dose 1 to before dose 2" group. Its efficacy in weeks 2-3 is better, and those end up in the "After dose 1" group.
Hold the phone, what's that second result line about? Result CI from between first shot and second shot is only 52%, with a wicked-big range of 29-64. I can't think of a good reason that this would vary so widely from the one treatment only study. In fact, they should be pretty close, with the after first-shot, before second shot being more likely to be HIGHER CI% than one shot only, because less elapsed time to get SARS CoV-2 before booster. Is this just because data are not robust? 21,000 is a pretty good number, but what was the N for just the "in between shots" treatment?
Please don't make me read the whole study :(
Heh. Honestly, the confidence interval is huge but I think the bigger question is durability. The immunologists I spoke with were a lot more confident about the first month than what happens in the longer run.
I'll have to take their word for it. You showed a lot of guts to run that opinion in the NYT, doesn't seem to have done any harm. Looks like the 2 dose decision was made very early on, so that's what we are stuck with, research wise. J&J is next in line, and if they get approved (when), that will really accelerate the number of people getting doses. What bothers me is that they are holding back half the Phizer/Moderna doses for the first tranche, to assure a second dose. This is where I would like to see a more adventurous use of resources to 'double' the # of recipients. So much worry about exact timing of 2nd dose! Why?
On this topic, I'm planning on listening to what the epidemiologists suggest. At the moment, it doesn't seem like there is sufficient peer reviewed opinion to suggest that the single dose route is appropriate.
Example: https://twitter.com/VirusesImmunity/status/1336323016737755142
She's an immunologist (even more appropriate!) but it doesn't answer the question at hand—the case for the booster is clear, but the trade-off isn't a pure immunology question because of the shortage. At some point, we will have data on single dose and if it is indeed reasonable and efficacious for at least a few months, we will be able to go back and calculate excess deaths. Hence the thorniness.
Another challenge would to be how a single-dose option might be received in under-served (and/or historically badly-served) communities. Does the single-dose option look like an abusive cost-cutting strategy, wrapped up in some marketing re: more lives saved?* Some folks get the low cost option w/higher risk; others get the high quality/cost full treatment w/lower risk? (Similar to the problem of Trump-adjacent people getting treatments not generally available.)
There are some communities who are vaccine hesitant due to historical medical missteps and worse. There are ethical issues re: treating them fairly and forthrightly. I'd want to get input from key players in those communities re: the single-dose idea. Regaining their trust need to be part of the equation.
* I know this is not where you are coming from. It is however, a possible interpretation for someone who lacks trust in the medical system.
Agree and absolutely. Hence a trial. Personally, I'd take the single dose without a trial but I'm not high risk.
(And yes, my bad re: her quals. A carry over from my Twitter list name for epi and immu types.)
It actually strengthens your point, that's the only reason I pointed it out. :-D
Yes, it's a thorny case of how to reduce overall harm.
I'm open to the idea, but after all of the unfortunate Covid-related political decisions (hydroxychloroquine, etc), I'm firmly in the camp of go slow and behave conservatively. Esp when in a situation w/so many unknowns.