I wonder if someone can put me straight here?
DS has been running a series of articles identifying the waning efficacy of the vaccines. The last one said something like "vaccine efficacy now minus 109%".
I have two problems with this, probably both due to my own limited brainpower.
First, how is it ever sensible to talk about vaccine efficacy as being "minus" anything? I appreciate that this was an attempt at comparing the vaccinated with the unvaccinated. But realistically (and unless the vaccine is actively giving you an illness or killing all its recipients) a vaccine either works as intended (i.e. reduces/eliminates death/serious illness) or it does not. It will not work in 100% of cases, so I can understand why efficacy might be below 100%. And if it was totally useless, and didn't work at all, I can understand efficacy being 0%. But surely it is a an unhelpful and possibly misleading manipulation via comparison to create such striking "minus" figures, particularly given the extremely high % of vaccinated compared to the unvaccinated? The impression given is that you are more likely to be infected if you are vaccinated, which is simply not true.
Secondly, all this talk of steeply declining vaccine efficacy seems to me to be a grand exercise in missing the point. Britain has had an average of 30,000-40,000 (roughly) "cases" a day for months now and yet hospital admissions and deaths remain stubbornly flat. The vaccines obviously work by reference to the only important measure, namely the ability to limit serious illness and death. Therefore, running articles which splash with headings and phrases such as latest figures show "an unadjusted vaccine effectiveness of minus 109%" when this only applies to protection against infection is, in my view, unhelpful to open debate.
I haven't missed the fact that there is a line later on in the report clarifying the distinction, but that only shows if you click "Read More", and not everyone has time every day to do that.
I am a committed lockdown/vaccine passport etc sceptic and firmly believe everyone should be free to choose whether to take the vaccine or not. But I do think that articles on this site should avoid the pitfalls seen in the MSM of glossing over unwelcome facts to make a preferred point.
I wonder if someone can put me straight here?
DS has been running a series of articles identifying the waning efficacy of the vaccines. The last one said something like "vaccine efficacy now minus 109%".
I have two problems with this, probably both due to my own limited brainpower.
First, how is it ever sensible to talk about vaccine efficacy as being "minus" anything? I
...
Secondly, all this talk of steeply declining vaccine efficacy seems to me to be a grand
..
Let see. The only article I could find with "negative" percentages were ones dealing with relative probabilities. Of vaccinated v unvaccinated groups. If the efficacy provided protection (lower infection rate) then the vaccinated population had a positive percentage probability. If the vaccinated group had a higher infection rate then the vaccinated population had a negative percentage probability. So -86% would mean the vaccinated group had a 86% higher infection rate than the unvaccinated. As I read the article. Thats all.
As for "cases" versus hospital admissions it means nothing regarding any purported efficacy of the vaccines. For that you need to compare actual active infection rates in the two groups, not clinically invalid screening test "cases". Anyway, the vaccine types used and the target is such than any efficacy will wear off in a matter of month. If there is any protection for the high risk groups. Which give the research from the flu shots over the years, is unlikely.
Plus SARs 2 seems to be like SARs 1 in 2003, mostly a hospital / health care facility acquired infection. So general population infection rates not that important.
So a causal reading of low resolution / low quality tells you nothing. In the bigger scheme of things. But annual death rates / 100K have not budged much in the last year. For the 10 year series. Which is a first for a supposed pandemic. Even H1N1-09 gave two nice bumps in 2010 and 2012. And Hong Kong flu in 1968 and Asian flu in 1958 had very noticeable spikes in mortality rates.
As for efficacy, think placebo. Which is all it will prove to be. Epidemiological speaking.
"...glossing over unwelcome facts to make a preferred point..."
In order to understand a vaccine efficacy given as -109% it is necessary to understand basic arithmetic and apply it to the standard formula used to calculate vaccine efficacy. The standard formula is given on this WHO web page. I have copied the key graphic and attached it to this post. I hope it is visible at the end of the post. As you can see from the WHO graphic the percentage efficacy is calculated as follows -
((no. of cases in placebo group) - (no. of cases in vaccinated group))/100
In the WHO example this is (10 - 2)/100 = 80% efficacy
To reach a figure of -109% is simple. Consider a much larger trial in which 200,000 people are involved. 100,000 are given the placebo and 100,000 are given the vaccine. OK so far? After testing we discover that 50,000 cases are found in the placebo group BUT 60,900 are found in the vaccinated group. Inserting these figures into the formula we have -
(50,000 - 60,900)/100
This gives -
-10,900/100
Carrying out the division gives a vaccine efficacy of -109%. The figure makes perfect sense & is consistent with current medical usage of the term "vaccine efficacy". Apparently it is you who is "...glossing over unwelcome facts to make a preferred point..."
thinksaboutit claims "You correctly point out that -109% is obviously nonsense." But thinksaboutit hasn't thought about it. I'm afraid he is, as ever, wrong.
"...To reach a figure of -109% is simple..."
Well, that'll teach me. As people have pointed out it's wrong. I misunderstood the basic formula. I'll try again. The percentage efficacy is -
((no. of cases in the placebo group) - (no. of cases in the vaccinated group)) /(no. of cases in the placebo group) x 100%
Applying this to produce a vaccine efficacy of -109% is still simple. Honest. Using 'p' as the number of cases in the placebo group & 'v' as the number in the vaccinated group the formula looks like this -
((p-v)/p) x 100%
Consider a trial with 300,000 participants. 150,000 in the placebo group & 150,000 in the vaccinated group. If we find the number of cases in the placebo group to be 50,000 then the number in the vaccinated group has to be 104,500. This gives the following -
((50,000-104,500)/50,000) x 100%
which is
((-54,500)/50,000) x 100%
which is
(-1.09) x 100%
which is
-109%
These figures work and make sense. I invite you to prove me wrong, again.
@halfhearted Thanks for your detailed response. I am sure all your calculations are correct, but I feel a little bit like giving Samuel Johnson's response to Bishop Berkeley's attempts to prove the non-existence of matter and that everything in the universe is ideal: Johnson gave a nearby rock an almighty kick and said "I refute it thus".
In the real world, vaccines either work, or they do not. That is to say, they either give some actual protection against disease, or they give none. That basic fact (i.e. does it give protection against disease?) cannot change if you compare vaccinated with unvaccinated no matter how detailed the calculation. Vaccines that give some protection against disease cannot have an efficacy of minus anything. And my objection to the article was the suggestion otherwise by way of a convoluted and unhelpful comparative calculation.
The big change from last year's figures is the vaccine intervention and that seems to me to be the obvious explanation for the current state of play. If there are other explanations for the lower hospitalisation/death rates (other than ones I've suggested) please do name them.
@shotclog in the pfizer trial, 44k people were given the vaccine or the placebo, spilt half and half. people were then sent away to get sick, 162 people in the placebo group got sick in the period, while 8 in the vaccine group got sick, indicating the placebo was 20 times less effective than the vaccine.That's why people take the pfizer vaccine.
You correctly point out that -109% is obviously nonsense.
But the aim of such articles is to influence the reader towards the opinion of the author. We can only assume it's just part of an anti-vaxx stance, which are powered by misinformation.
Hello there,
I'm no expert, but could the strange negative value of efficacy be due to the misguiding way the vaccine efficacy is calculated? Are you familiar with RRR (Relative Risk Reduction) and ARR (Absolute Risk Reduction)?
From what I understand, Pfizer started off with 95% (RRR) efficacy which dropped to around 50% due to the Israeli vaccination outcomes.
I pasted in a spreadsheet I made a few weeks ago to try to understand the weird way this is all calculated.
I hope it helps in some way.
| Pfizer | Placebo | Oxford AZ | Placebo | Ivermectin | Placebo | |||
| All Participants | 17411 | 17511 | 13934 | 13883 | 919 | 505 | ||
| Didnt Catch | 17403 | 17349 | 13923 | 13698 | 919 | 257 | ||
| Caught Covid | 8 | 162 | 11 | 185 | 0 | 248 | ||
| % who caught | 0.05% | 0.93% | 0.08% | 1.33% | 0.00% | 49.11% | ||
| % who didnt | 99.95% | 99.07% | 99.92% | 98.67% | 100.00% | 50.89% | ||
| extrap to 100k | 46 | 925 | 79 | 1333 | 0 | 49109 | ||
| Relative Risk (RRR) | 95% | 94% | 100% | |||||
| Absolute Risk (ARR) | 0.88% | 1.25% | 49.11% |
This doesn't explain a negative efficacy at all.
Negative efficacy would mean the treatment increases the incidence, relative to the placebo.
Looking at your numbers, you take the trouble to include ivermectin, which is considered a feasible treatment by some (as yet unproven in any meaningful trials).
Your numbers seem to suggest a 100% efficacy for prevention, which is obviously not true. Then compared to the inexplicable 49% infection rate for placebo in this sample. Doesn't compare well to all the other placebo groups. About 50x worse!!!!!!!!!!!!!!!!!!!
Can you advise where you got these numbers from?
Yes my apologies, please disregard the Ivermectin figures from the above spreadsheet, I posted in haste this morning!
These are the Ivermectin figures I collected a few weeks back. I understand some stuidies have been retracted since that time though:
| https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8088823/" }" data-sheets-hyperlink="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8088823/"> | |||||
| Meta Analysis | |||||
| Ivermectin | Control | ||||
| Caught | Total | Caught | Total | ||
| Behera | 15 | 91 | 171 | 281 | |
| Carvallo-1 | 0 | 131 | 11 | 98 | |
| Carvallo-2 | 0 | 788 | 237 | 407 | |
| Alam | 4 | 58 | 44 | 60 | |
| Elgazzar | 2 | 100 | 10 | 100 | |
| Shournan | 15 | 203 | 59 | 101 | |
| Chala | 4 | 117 | 25 | 117 | |
| Total | 40 | 1488 | 557 | 1164 | |
| % who caught | 2.7% | 47.9% | |||
| % who didnt | 97.3% | 52.1% | |||
| extrap to 100k | 2688 | 47852 | |||
| Relative Risk (RRR) | 93% | ||||
| Absolute Risk (ARR) | 45.2% |
The Pfizer and Moderna figures were taken from page 24 of the EUA.
