0

According to the CDC's page on the safety of J&J vaccines, the reported efficacy rate of the vaccine is a staggeringly low 66.3%.

That said, the government is still promoting and touting the J&J vaccine as a reliable and effective way to protect oneself against the effects of the COVID-19 virus.

In a report published just yesterday, the CDC published an official study on 32,867 COVID vaccinated people and found numbers that closely matched those of the clinical trials mentioned above, in that Moderna is 95% effective at preventing hospitalization, followed by Pfizer at 80% and J&J at 60%.

So if the numbers are correct, why is J&J still touted as an effective protection against COVID-19 still? 60% isn't exactly the greatest confidence booster -- that's almost at the level of a coin flip. How can the J&J vaccine shot be effective against COVID-19 if it only has a ~60% efficacy rate? Pfizer/BioNTech and Moderna's vaccines offer a much more solid guarantee, but 60% for J&J?

How can 60% be considered effective?

yuritsuki
  • 1,965
  • 2
  • 15
  • 25
  • 6
    This is a more suitable Q to ask on https://medicalsciences.stackexchange.com/questions Your're not really disputing any data/figures, just quibbling with how the CDC phrases their findings. I'd also add that efficacy for mRNA vaccines seems to drop faster over time compared to the AD-vectored ones, from other recent studies (including some by the CDC), see e.g. https://medicalsciences.stackexchange.com/questions/28894/is-there-uk-data-on-the-reduction-in-vaccine-effectiveness-for-people-who-were-v – Fizz Sep 11 '21 at 06:59
  • 8
    You really need to distinguish effectiveness against infection and against serious illness/hospitalization and death. Those are very different numbers, there is not a single effectiveness number for the vaccines – Mad Scientist Sep 11 '21 at 08:11
  • Not sure why you think something being 60% better is the same as "maybe better, maybe not at all." If I have unvaccinated 500 people, and 100 can be expected to get covid, and with a vaccine that number is 40/500, that's effective and helpful, overall. You're saving 60% of the potentially would-be infected people from getting it. – PoloHoleSet Oct 25 '21 at 21:45
  • @PoloHoleSet Because when there are alternatives that could save 90% or even 99%, why would you EVER settle for 60%? – yuritsuki Oct 26 '21 at 21:10
  • @yuritsuki the question of why would you chose JJ over the others is a completely different question. I will try to shortly answer it now, but feel free to ask a different question for a more detailed anwer. Convenience (1 shot means getting the QR code sooner), availability (some only have access to JJ), price (10 times more vaccinated people with 60% is better than a tenth of the people vaccinated with 99%). – Andrei Nov 02 '21 at 18:09

1 Answers1

17

60% isn't exactly the greatest confidence booster -- that's almost at the level of a coin flip.

That's actually not quite so.

So, for example, let’s imagine a vaccine with a proven efficacy of 80%. This means that – out of the people in the clinical trial – those who received the vaccine were at a 80% lower risk of developing disease than the group who received the placebo. This is calculated by comparing the number of cases of disease in the vaccinated group versus the placebo group. An efficacy of 80% does not mean that 20% of the vaccinated group will become ill.

it does mean that in a vaccinated population, 80% fewer people will contract the disease when they come in contact with the virus

VE (vaccine efficacy) is calculated as (ARU - ARV) / ARU, where ARU and ARV are the attack rates among unvaccinated and vaccinated groups, respectively.

If the ARU is say 0.7, i.e. without a vaccine you get ill 70% "of the time" (more correctly 70% of the population would--something like this was estimated early in the pandemic) then a VE of 0.6 means that that ARV is just 0.28, i.e. only a bit more than a quarter of the vaccinated would get infected. And that's much better than a coin flip. So ARV, which is really what you want to compare with the coin flip, really depends both on the ARU and VE.

Also, the VE numbers you're talking about concern hospitalization... and the ARU for that is pretty far from a coin flip to begin with--something like 5% (based on an infection hospitalization rate of 7% among the infected, and assuming 70% of the pop gets infected). So, in this latter case, comparing two numbers (ARU, ARV) that are far from a coin flip... with a coin flip is not exactly insightful. ARV would be 0.02 in this case, i.e. your "hospitalization chance" drops from 5% to 2% when vaccinated, under these assumptions. That may not look like much (absolute) difference for individual protection, but it does mean needing less than half as many hospital beds etc.

Somewhat o/t, but in a 3rd world country (that can't afford mRNA vaccines) even the J&J vaccine could make quite a difference as anyone needing hospital care and not getting it at all is fairly likely to die; see what shortages of oxygen have caused in such counties.

And hospitalization isn't the only thing worth considering. Deaths were not considered in that CDC study you're quoting from, but the J&J vaccine seems to do better in that regard:

in early August, a clinical trial that followed nearly 500,000 health care workers in South Africa found that the J&J vaccine was 71% effective against hospitalization and 95% effective against death due to the delta variant.

I'm not going to comment on how the CDC chooses to phrase their findings when comparing vaccines, but I'll just add that efficiency at single point in time is not the whole story.

By the way, there's a study based on a multiple-choice question in a sample from the general public that finds that the concept/measure of VE is highly unintuitive: only 3% of those questioned picked the correct definition; 64% confused it with (1-ARV) i.e. the percentage of individuals who do not develop Covid-19 among those vaccinated.

I wasn't too inclined to get into relative efficacy (of one vaccine vs another) here--since no numbers were disputed just the wording, but--in the meantime--the CDC & FDA have pretty much admitted that J&J isn't as effective as the mRNA vaccines, at least in terms of recommending a booster sooner for J&J (after 2 months instead of after 6 as for mRNA ones) and for everyone over 18 for J&J as opposed to select categories of risk for the mRNA vaccines; FDA's statement:

The use of a single booster dose of the Moderna COVID-19 Vaccine that may be administered at least 6 months after completion of the primary series to individuals:

  • 65 years of age and older
  • 18 through 64 years of age at high risk of severe COVID-19
  • 18 through 64 years of age with frequent institutional or occupational exposure to SARS-CoV-2

The use of a single booster dose of the Janssen (Johnson and Johnson) COVID-19 Vaccine may be administered at least 2 months after completion of the single-dose primary regimen to individuals 18 years of age and older.

The CDC's statement is almost identical except for some minor wording differences:

For individuals who received a Pfizer-BioNTech or Moderna COVID-19 vaccine, the following groups are eligible for a booster shot at 6 months or more after their initial series:

  • 65 years and older
  • Age 18+ who live in long-term care settings
  • Age 18+ who have underlying medical conditions
  • Age 18+ who work or live in high-risk settings

For the nearly 15 million people who got the Johnson & Johnson COVID-19 vaccine, booster shots are also recommended for those who are 18 and older and who were vaccinated two or more months ago.

There's more (interesting) background on how this decision was reached reported on NPR on Oct 21; that story also notes that the recommendation for Moderna matched that for Pfizer.

And this is somewhat obvious given the technological parallel that J&J's vaccine has with AstraZeneca or Sputnik V used in other countries, but when this latter FDA & CDC announcement was in preparation news was that "Fauci said FDA panel's recommendation shows J&J vaccine 'should have been a two-dose vaccine to begin with'".

Fizz
  • 57,051
  • 18
  • 175
  • 291
  • 1
    But your numbers *do* indicate that the coin flip metaphor is not too far from correct: If 5% of the population "win" a hospitalization ticket and - thanks to vaccination - use a coin flip to decide whether they actually need it, we end up with 2.5% hospitalization (instead of 2%, but then again, a 50% coin flip is not a 60% thing) – Hagen von Eitzen Sep 12 '21 at 05:47
  • 7
    @HagenvonEitzen: but not being vaccinated has 100% chance in that conditioned case. So it's not clear why the coin flip is the relevant comparison. You could say that it goes from a "sure thing" to something like a coin flip (even in that conditioned case), but even that's an improvement. I mean, you do have a point, but it's coin-flip vs 100%. – Fizz Sep 12 '21 at 11:29
  • These fancy theoretical models are all well in line with official propaganda. However, "100%"? What's that to mean? There is pre-existing immunity, innate and cross, plus recovered & young people, risk is not evenly spread but stratified, IFR quite low, ARI spread in waves, across subgroups, never to all, and finally ARR for J&J is abysmal _0_.93%, according to trial data. If you want to praise J&J, say it's better than Pfizer at _0_,84% ARR! But on a population level, both failed spectacularly, as ARR already predicted… We just have spike-escape breeding grounds in an immune forest fire. – LangLаngС Oct 18 '21 at 12:24
  • In this "conditioned case" you reference in comment, is that to say that this 60% figure is those spared hospitalization among those who actually become ill, or does it calculate into it the vaccine's protection against getting ill in the first place? I'm trying to follow the line of sub-groups that leads to the benefactors, but it's a bit confusing. –  Oct 18 '21 at 14:00
  • I understand such a figure might not exist, but putting it in terms of a population of 1000, noting how many we expect to become ill, hospitalized, and killed (assuming only x-vaccine is used) would be far easier to understand. –  Oct 18 '21 at 14:02
  • 2
    @LangLаngС: In how far does your comment contain suggestions to improve this answer? – Schmuddi Oct 18 '21 at 14:12
  • 1
    @Schmuddi 1. I'd suggest to ditch the theoretical calculations, they're irrelevant ('infections' are no end point, the attack rates and percentages dystopian fantasy) 2. I'd suggest to use the actual trial data (although they're all prematurely abandoned, not least by destroying the control groups) 3. to hilight these limitations 4. to compare with observational results waning from marginal start 5. to introduce ARR & NNTT/NNTV for comparison, those are required for interpretation, always // In this way you may be able to see how neither Pfizer nor J&J _can_ deliver what is promised. – LangLаngС Oct 18 '21 at 14:26
  • 4
    @LangLаngС: like it or not VE is the measure used in reporting such results. If it's a "fantasy" to you, fine, go ahead and revolutionize the field. Note that VE is just a relative risk ratio, fundamentally, so it's not that unusual/different from what's reported elsewhere in medicine, although there are indeed debates whether one should accompany relative risk measures by absolute ones. – Fizz Oct 18 '21 at 14:43
  • Me & revolutionize anything? Would be late to the game 10.1016/S2666-5247(21)00069-0 10.1101/2021.06.28.21259039 – LangLаngС Oct 18 '21 at 15:11
  • https://blogs.bmj.com/bmj/2021/01/04/peter-doshi-pfizer-and-modernas-95-effective-vaccines-we-need-more-details-and-the-raw-data/ 10.4103/2229-3485.173773 10.1093/ndt/gfw465 10.3390/medicina57030199 and hundreds other. // Failure to adequately communicate the so called VE _together_ with the other measures — and how they are by no means fixed but a result of complex interactions that change –was a disgusting salesman tactic to prop up the dressing window. – LangLаngС Oct 18 '21 at 15:21
  • 1
    @LangLаngС: ["Doshi had relied on some very weird arguments."](https://absolutelymaybe.plos.org/2021/08/31/this-waning-immunity-argument-against-the-fdas-covid-vaccine-approval-is-a-scientific-quagmire/) And that's an actual quote, unlike your "//" statement. – Fizz Oct 18 '21 at 15:33
  • Telling how that picks the straw by its horns? What you say 'I may revolutionize' is already b/c "both doctors and lay people tend to overestimate the effect when it is presented in terms of relative risk" ~ ["endorsed by over 50% of the core medical journals listed in the Abridged Index Medicus on PubMed"](http://www.consort-statement.org/about-consort/endorsers1). Only that the pharma-shills for vax gleefully ignore that as readily as this A does. Take Cochrane's Gøtzsche for a nice quote: "Significant results in abstracts are common but should generally be disbelieved." BMJ 2006;333:231-4." – LangLаngС Oct 18 '21 at 16:36
  • 5
    @LangLаngС: What do (marginal) p-values ([Gøtzsche's paper](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1523498/) concerns in particular papers reporting p-values between 0.04 and 0.06) have to do with VE? Please stop posting more and more tangential material under this answer. (Aside: I couldn't find the original J&J paper in pinch, but for their booster results, they give 95% CI for VE but not the much chastised p-value https://www.jnj.com/johnson-johnson-announces-real-world-evidence-and-phase-3-data-confirming-strong-and-long-lasting-protection-of-single-shot-covid-19-vaccine-in-the-u-s – Fizz Oct 18 '21 at 16:41
  • The idea that only 3% of those polled understood vaccine-efficiency seemed pretty extreme, so I checked out [that study](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8314794). For that **"_Experiment 1_"**, the study notes that the survey was done online for participants paid for their time, and: **"_There were no time limits on task completion, and the average response time was less than 2 min._"**. Usually, averages are higher than medians in scenarios like that, and the multi-choice definitions were really wordy, so I'd be suspicious that many survey-takers just randomly responded. – Nat Oct 19 '21 at 13:13
  • Apparently those who responded received a flat rate of **"_0.63 British pounds_"**, without apparent reason for them to try to untangle the verbose alternative definitions. A better study probably should've given participants a hypothetical scenario and asked them to calculate the vaccine-efficiency. For example, "_40% of unvaccinated were infected, vs. 10% of vaccinated. What's the vaccine-efficiency?_". Something like that should be simple enough to not annoy participants into random-selection (or at least less so). – Nat Oct 19 '21 at 13:15
  • Okay, (Section 3.1.4) of that study reports that, when survey-takers were asked to give an open-ended response (rather than select a multiple-choice option), they judged that only about 1% of respondents got it about right. _However_, after reading through their interpretations of popular descriptions, it sounds like the authors of the study are incorrectly classifying some definitions as wrong. For example, the study incorrectly criticizes another study's definition, and then applies the same criticism to the bulk of the free-responses. They may be attacking a straw-man in free-response. – Nat Oct 19 '21 at 13:44
  • (Specifically, some folks may describe 95% vaccine-effectiveness as providing 95% protection -- which is true! The authors seem to think that description's false under a straw-man interpretation of it.) – Nat Oct 19 '21 at 13:48
  • 2
    I don't see the difference here. 60% effectiveness means if you would have got the disease, you now flip a coin with 60% heads, and if it comes up heads, you are protected. The other 40% of the time, the vaccine did not protect you. If you are one of the % of people who would not have gotten the disease, then we don't know whether the vaccine would protect you, but we can infer it's the same chance as among the people who would have. This is the obvious way to measure effectiveness of an intervention against some danger. – user253751 Oct 25 '21 at 10:40