4 ms·

They do not have different measures of eficacy. AZ did more testing to look at symptomatic infections but that was not used in the primary efficacy calculation.

by m-ee 6y ago

They do not have different measures of eficacy. AZ did more testing to look at symptomatic infections but that was not used in the primary efficacy calculation. https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(20)32661-1/fulltext https://www.thelancet.com/journals/lancet/article/PIIS0140-6...

boulos 6y ago
I believe AZ was doing weekly testing while Moderna (maybe?) was doing “test if symptomatic”. That’s a huge difference.
- m-ee 6y ago
  If they didn't have symptoms, they weren't counted in the headline efficacy numbers we're talking about. "The primary outcome was virologically confirmed, symptomatic COVID-19, defined as a NAAT-positive swab combined with at least one qualifying symptom (fever ≥37·8°C, cough, shortness of breath, or anosmia or ageusia)."
mlyle 6y ago
Just a nit: they don't have exactly the same outcome measure, definition of "severe", "symptomatic", etc. Nor were they conducted in the same, time, or the exactly same demographics. Nor did they use the same protocol: testing weekly with a swab and detecting and informing the participant of a positive status will get that person to report symptoms to the trial more often than if you don't. They use similar measures of efficacy but they're not the same thing and are not perfectly comparable. Still, it seems very likely that the mRNA vaccines are much better than AZ.
- m-ee 6y ago
  You’re absolutely right, but I didn’t want the overall message to get lost in the details. I’ve been seeing people parrot the “AZ is actually much better than it looks because they included asymptomatic patients” for months now and it’s just incorrect.
  - boulos 6y ago
    I’m honestly confused! Can you point me to your source for the excerpts in your sibling comments? Edit: reading your lancet link now (sorry about that). My understanding from reading the description of the Moderna results was “XX / YYY people developed covid; but we didn’t test everyone in the YYY set” (so people who definitely got it / went to a hospital / whatever are in XX, but asymptomatic cases may not be). Agreed? My understanding of the AZ studies (despite all their other mistakes) is that participants were being tested weekly, regardless of symptoms. Is that incorrect, or are you suggesting something else? I definitely don’t want to repeating misinformation. Second Edit: after looking at table 2 in [1] and comparing the results to the common headline “about 60% effective” [2], it seems like media are reporting the “two standard doses” results. That might be what some folks end up doing, and it’s regretful that the study design went all over the place. Either way, thanks for the nuance! I will update my future comments to note that the confidence intervals including the asymptomatic results are wildly large, such that they need serious discounting. [1] https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(20)32661-1/fulltext https://www.thelancet.com/journals/lancet/article/PIIS0140-6... [2] https://www.bbc.com/news/world-europe-55862233 https://www.bbc.com/news/world-europe-55862233
    - m-ee 6y ago
      A subset of people in the AZ study were tested weekly. However the primary endpoint was prevention of symptomatic disease. For someone to count in the top line efficacy number they would have to show symptoms and test positive on PCR. This is roughly equivalent to what the other studies did, although exact definition of "symptomatic" and how they monitored that can be slightly different. There's a common perception I've seen a lot that the AZ numbers are sandbagged because it included asymptomatic cases but that's not the case. If someone never developed symptoms they don't factor in to the 60% efficacy number. As the other commenter pointed it may be possible that someone had mild symptoms they were ignoring then tested positive on the weekly screening and said "oh yeah I do have a sore throat" but the numbers should still be roughly comparable. It's not worth diving into 94% vs 95% efficacy but I think we can be pretty confident that the 95% vaccine is more effective than the 60% vaccine. This is also muddied because the AZ trials were a bit of a mess with dosing error and varying times between shots. So now we have a number of different subgroups with their own statistics but not a big enough sample size to say anything confidently about half dose vs full dose or 2 weeks vs 12 weeks. The confidence intervals are a mile wide, but they still get talked about (and even drive policy) adding a lot of noise to the discussion about how effective the vaccine really was.

by m-ee 6y ago

boulos 6y ago

I believe AZ was doing weekly testing while Moderna (maybe?) was doing “test if symptomatic”. That’s a huge difference.

m-ee 6y ago
If they didn't have symptoms, they weren't counted in the headline efficacy numbers we're talking about. "The primary outcome was virologically confirmed, symptomatic COVID-19, defined as a NAAT-positive swab combined with at least one qualifying symptom (fever ≥37·8°C, cough, shortness of breath, or anosmia or ageusia)."

mlyle 6y ago

Just a nit: they don't have exactly the same outcome measure, definition of "severe", "symptomatic", etc. Nor were they conducted in the same, time, or the exactly same demographics. Nor did they use the same protocol: testing weekly with a swab and detecting and informing the participant of a positive status will get that person to report symptoms to the trial more often than if you don't. They use similar measures of efficacy but they're not the same thing and are not perfectly comparable. Still, it seems very likely that the mRNA vaccines are much better than AZ.

m-ee 6y ago
You’re absolutely right, but I didn’t want the overall message to get lost in the details. I’ve been seeing people parrot the “AZ is actually much better than it looks because they included asymptomatic patients” for months now and it’s just incorrect.
- boulos 6y ago
  I’m honestly confused! Can you point me to your source for the excerpts in your sibling comments? Edit: reading your lancet link now (sorry about that). My understanding from reading the description of the Moderna results was “XX / YYY people developed covid; but we didn’t test everyone in the YYY set” (so people who definitely got it / went to a hospital / whatever are in XX, but asymptomatic cases may not be). Agreed? My understanding of the AZ studies (despite all their other mistakes) is that participants were being tested weekly, regardless of symptoms. Is that incorrect, or are you suggesting something else? I definitely don’t want to repeating misinformation. Second Edit: after looking at table 2 in [1] and comparing the results to the common headline “about 60% effective” [2], it seems like media are reporting the “two standard doses” results. That might be what some folks end up doing, and it’s regretful that the study design went all over the place. Either way, thanks for the nuance! I will update my future comments to note that the confidence intervals including the asymptomatic results are wildly large, such that they need serious discounting. [1] https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(20)32661-1/fulltext https://www.thelancet.com/journals/lancet/article/PIIS0140-6... [2] https://www.bbc.com/news/world-europe-55862233 https://www.bbc.com/news/world-europe-55862233
  - m-ee 6y ago
    A subset of people in the AZ study were tested weekly. However the primary endpoint was prevention of symptomatic disease. For someone to count in the top line efficacy number they would have to show symptoms and test positive on PCR. This is roughly equivalent to what the other studies did, although exact definition of "symptomatic" and how they monitored that can be slightly different. There's a common perception I've seen a lot that the AZ numbers are sandbagged because it included asymptomatic cases but that's not the case. If someone never developed symptoms they don't factor in to the 60% efficacy number. As the other commenter pointed it may be possible that someone had mild symptoms they were ignoring then tested positive on the weekly screening and said "oh yeah I do have a sore throat" but the numbers should still be roughly comparable. It's not worth diving into 94% vs 95% efficacy but I think we can be pretty confident that the 95% vaccine is more effective than the 60% vaccine. This is also muddied because the AZ trials were a bit of a mess with dosing error and varying times between shots. So now we have a number of different subgroups with their own statistics but not a big enough sample size to say anything confidently about half dose vs full dose or 2 weeks vs 12 weeks. The confidence intervals are a mile wide, but they still get talked about (and even drive policy) adding a lot of noise to the discussion about how effective the vaccine really was.