Vaccine Effectiveness in Ireland

Cuddlesworth

https://www.boards.ie/discussion/comment/117807868#Comment_117807868

When you change values in a equation, you get different results.

Do you have the actual percentage values of the vaccination status for hospital admissions over a 7 day period, not the percentage of vaccinated people in hospital at a given time?

Mr.StRiPe

https://www.boards.ie/discussion/comment/117808001#Comment_117808001

When you change values in a equation, you get different results.

Great so you agree with that too

The next question is do you agree my spreadsheet accuratly reflects the model professor Nolan presented with the 5 variable inputs we agreed on previously?

Unicorn Milk Latte

https://www.boards.ie/discussion/comment/117807568#Comment_117807568

With 80%+ of the population vaccinated

As of today, in Ireland 65.2 percent of the population are fully vaccinated.

The 80%+ number you mention means you probably mixed up 'population' with the 18+ age group of the population. 85.5% of over-18s fully vaccinated as of today.

Mr.StRiPe

https://www.boards.ie/discussion/comment/117808968#Comment_117808968

As of today, in Ireland 65.2 percent of the population are fully vaccinated

This is something I'd also like to bring into the discussion but first we need to establish if we agree on what the model means and if my spreadsheet accuratly reflects this otherwise we can't move on to putting the real data into the model.

In the model professor Nolan posted he qualifies that he is talking about 70% of the adult population protected (two weeks after completing their regimen) so this is what I have gone with

When I referenced the 80%+ I was referencing myself as I am an adult. I know it has been pointed out I could be making that up and there is no reason to believe me as some random person on the internet which is true too.

However, if you look at the real data available you could quite accuratly come to the coclusion that I am indeed an adult in fact you could get an even more accurate estimate by looking at my boards join date which is July 1999 which would make me at least 22.

You could go further looking up the history of boards.ie which used to be a quake forum to predict I was likely involved at that point and that was comprised of people in irish colleges who were playing quake on the LAN's so would likely mean I was an adult when I joined and the main 2 colleges involved in this at the time were UL and TCD etc etc This also means the older I am the more likely I am vaccinated etc.

This is an illustration of how real data can be used to get an accurate estimation of the reality it exists in.

I agree you can't be 100% sure based on the data as we could be misinterpreting and adding our own bias to it hence we need to flesh it out with discussion where people are willing to discuss and not let their own bias prevent them from properly engaging.

is_that_so

https://www.boards.ie/discussion/comment/117808968#Comment_117808968

With the 12-15 group we might actually hit 80%+ fully vaccinated. We are currently heading for close to 75% with the adult numbers, and that cohort is about 5% of the population.

Mr.StRiPe

https://www.boards.ie/discussion/comment/117803278#Comment_117803278

I'm talking about drawing conclusions about the efficacy of vaccines from 'x% of hospitalised are vaccinated', because that is misleading when the number of hospitalised overall is small, while the number of overall vaccinated in the population is high.

The model mathematically models the relationship between the % of hospitial admissions and the effectiveness of the vaccine in preventing this based on the % of the population fully vaccinated.

This means it doesn't matter if the number of hospitalised overall is a small or large number what matters is what % of hospital admissions that are fully vaccinated.

As we move though different waves of the pandemic when the wave is increasing we get larger numbers in hospital week on week and when the wave is decreasing we get less numbers week on week and inbetween we get a more consistent level. However, at all times the % of those addmitted to hospital that are fully vaccinated should match, on average, the % of fully vaccinated admissions in the model for the % of the population vaccinated at that point.

If there is a signifficant difference between the real % of admissions in fully vaccinated people and the % in the model this indicates the models assumption for % vaccine effectiveness in preventing sever disease requiring hospitalisation is incorrect. If the real admission % of fully vaccinated is lower than the equivalent % in the model it indicates the vaccince effectiveness is underestimated and if it is higher it indicates it is overestimated.

This is a mathematically sound assumption to make based on what the model represents.

Lumen

You're still not getting it.

Let me spell it out.

Consider the following hypothetical age-related hospitalisation risk and vaccine coverage.

• 50% of the population ("old") has a base hospitalisation risk of 5%, and 90% vaccine coverage.
• 50% of the population ("young") has a base hospitalisation risk of 1%, and 70% vaccine coverage.
• Vaccines reduce the risk of hospitalisation by 95% in both age groups.

Now, run the numbers for a population of 1,000,000.

CRAP MODEL EXCLUDING AGE-RELATED DIFFERENCES IN VACCINE COVERAGE AND HOSPITALISATION RISK

The population-wide hospitalisation risk is 3%, the population-wide vaccine coverage is 80%.

Of the 1m population, 800k are vaccinated, and of those, 1200 are in hospital. 200k are unvaccinated, and of those, 6000 are in hospital.

The admissions split between vaccinated and unvaccinated is 17:83.

BETTER MODEL INCLUDING AGE-RELATED DIFFERENCES IN VACCINE COVERAGE AND HOSPITALISATION RISK

Of the 500k "old" population, 450k are vaccinated, and of those, 1125 are admitted to hospital. 50k are unvaccinated, and of those 2,500 are admitted to hospital.

Of the 500k "young" population, 400k are vaccinated, and of those, 200 are admitted to hospital. 100k are unvaccinated, and of those 1,000 are admitted to hospital.

Across the whole population, there are 1,325 vaccinated and 3,500 unvaccinated in hospital.

The admissions split between vaccinated and unvaccinated is 27:73.

Conclusion: ignoring age-related differences in hospitalisation risk and vaccine coverage will make the vaccines look less effective than they are, because sub-populations at higher risk tend to have better vaccine coverage.

It doesn't particularly matter what the specific values are, as long as vaccine coverage is correlated with risk you will see this modelling bias.

I welcome a homework check if anyone can be bothered.

edit: updated due to terrible maths at first attempt

Ramasun

I'd imagine 99% of motorcycle road fatalities involved a rider wearing a helmet.

Mr.StRiPe

https://www.boards.ie/discussion/comment/117823358#Comment_117823358

I have no problem fleshing out your suggested models further however, I need to confirm first that you agree it is mathematically sound to assume like professor Nolans model if we put the real HSE data into your model and the breakthrough admission ratio and/or admission % does not match the expected model output this indicates the assumption on the effectiveness of the vaccine with regard to hospitalisations is incorrect? i.e. it could be either over or under estimated

Lumen

No, I do not agree.

I'm simply giving an example of how you cannot infer vaccine effectiveness from two populations which differ in more ways than their vaccination status.

This is why clinical trials involve careful selection of groups to ensure that they only differ in one way, and (where possible) blinding to avoid placebo effect, treatment bias etc.

You've attached far too much weight to a tweeted infographic.

Mr.StRiPe

https://www.boards.ie/discussion/comment/117825961#Comment_117825961

You clearly do not understand how professor Nolans model works so I suggest you have a closer look at it. Eitherway, you'll be glad to know that not only does his model cover the 2 groups you mentioned "old" and "young" it covers all age groups within the adult population of Ireland.

If you had extrapolated your models out fully with your "hypothetical age-related hospitalisation risk and vaccine coverage" assumptions you would have found that the % of fully vaccinated hospital admissions would have been the same in both models. This is because you said "Vaccines reduce the risk of hospitalisation by 95% in both age groups" which factors in the "base hospitalisation risk" you referenced for each group. Thus the % of fully vaccinated hospital admissions will be the same for each group as they are based on 95% reduction in hospitalisation when compared to the base hospitalisation for each group note I am talking about the % not the actual number which will be different in each group due to the base hospitalisation risk being different.

Let me spell it out

if we take a group of people all aged 71 (old)

• The unvaccinated in this group will be addmitted to hospital at the base hospitalisation risk % for this group
• The vaccinated in this group will be admitted to hospital at the base hospitalisation risk % for this group reduced by the reduction in hopsitalisation benifit of the vaccine i.e. 95% in your models
• The % difference in expected hospital admissions in this group for both vaccinated and unvaccincated is based purely on the vaccine effectiveness figure i.e. 95%

if we take a group of people all aged 36 (young)

• The unvaccinated in this group will be addmitted to hospital at the base hospitalisation risk % for this group
• The vaccinated in this group will be admitted to hospital at the base hospitalisation risk % for this group reduced by the reduction in hopsitalisation benifit of the vaccine i.e. 95% in your models
• The % difference in expected hospital admissions in this group for both vaccinated and unvaccincated is based purely on the vaccine effectiveness figure i.e. 95%

Pick whatever age group you wish and the outcome is the same i.e. the same % of fully vaccinated people get admitted to hospital in each age group as this is based on the 95% effectiveness rate of the vaccine in each group. Again I am talking about the % of fully vaccinated admissions not the number in each group as the number varies based on the base hospitalisation risk for each group. i.e. it will be higher in older groups

As the expected admission % of fully vaccinated is the same for all age groups it is the same for the full population hence if the real hospital admission % of fully vaccinated is lower than the % value referenced by the model it indicates the vaccine effectiveness is underestimated and and if it is higher it indicates it is overestimated.

If you still don't believe me extrapolate out your models and you will see it.

Lumen

One of us is spectacularly wrong.

This feels qualitatively a bit like the plane on a conveyor belt problem.

https://www.boards.ie/discussion/comment/117830734#Comment_117830734

It is quite stressful to be exposed to such a failure of basic logic.

Mr.StRiPe

https://www.boards.ie/discussion/comment/117831167#Comment_117831167

Perhaps once you have gotten over your stress you can come back and contribute to the discussion by explaining why you feel the logic in professor Nolans model is incorrect to have the vaccine effectiveness % in preventing hospitalisations directly reflect the hospital admission % for both the vaccinated and unvaccinated for the population in his model.

Mr.StRiPe

https://www.boards.ie/discussion/comment/117831128#Comment_117831128

You choose to add the conveyor belt to professor Nolans "CRAP MODEL" by introducing your own "BETTER MODEL" model.

If you actually extrapolate out your model fully you will find there is no conveyor belt at all as it will agree with Nolans model when it comes to the % of admissions of fully vaccinated people.

Mr.StRiPe

https://www.gov.il/BlobFolder/reports/vaccine-efficacy-safety-follow-up-committee/he/files_publications_corona_booster-27082021.pdf

This is the initial analysis of the data from Israel that shows how effective giving a booster dose of vaccine is at correcting the issue of the waning effectiveness of the Pfizer vaccine against infection 5+ months out from receiving the 2nd dose.

This is interesting as it shows the booster has the effect of boosting the vaccines protection back up against both the relative risk of confirmed infection and the relative risk of severe illness by roughly the same levels i.e. 11.4 fold decrease in the relative risk of confirmed infection and greater than 10 fold decrease in the relative risk of severe illness.

This is the first time I have seen data that shows not only is the protection against relative risk of severe illness waning it is waning at a comparable rate to the protection against the relative risk of confirmed infection. This also supports what professor Nolans model showed when the real HSE data was entered into it.

This report gives the example of 50% effectiveness of protection against infection after a waning period of 6+ months based on the general concenses of recent reports looking at this.

If we input 50% as the vaccine effectiveness symptomatic % into professor Nolans model and reduce the vaccine hospitalisation reduction % by 45% (95 down to 50) we get the following:

These effectinveness %'s puts us into the ballpark of the 40%+ breakthrough hospital admissions we saw with the real HSE data in Nolans model when 70-75% of the adult population were fully vaccinated.

The model suggests the % effictiveness against hospitalisation in Ireland is slightly overestimated (7%) on the week of 12th of August and underetimated (4%) the week of the 19th of August.

The HSE haven't released any breakthough data since the 19th of August so this is all we can go on at this stage.

amandstu

From CNN today

https://edition.cnn.com/2021/09/08/health/severe-breakthrough-cases-cdc-studies/index.html

"But in those rare cases when a fully vaccinated person gets infected, data suggests it's older adults and those with multiple underlying medical conditions who are most at risk of serious illness" (my bold)

I am unclear from this article how effective the vaccine is against actually becoming infected...

I was under the impression that the vaccine protected against severe symptoms rather than infection per se but this article appears to be saying that it protects very strongly against not only symptomatic illness but also asymptomatic illness.

Is that right?

If we have a group of people where one of the group is infectious then the likelihood is that those who are unvaccinated may become infected but those who are vaccinated are very unlikely to become so ?

And ,if they did become infected they are as infectious (if for a shorter length of time) as those who became infected and were also unvaccinated?

Mr.StRiPe

https://www.boards.ie/discussion/comment/117866256#Comment_117866256

I am unclear from this article how effective the vaccine is against actually becoming infected...

I was under the impression that the vaccine protected against severe symptoms rather than infection per se but this article appears to be saying that it protects very strongly against not only symptomatic illness but also asymptomatic illness.

Is that right?

The effectiveness of the vaccine is typically measured in the following 2 ways:

1. its protection against symptomatic infection
2. its protection against hospitalisation and severe disease

These effictiveness measures are linked in as far as a vaccinated person needs to have a breakthrough symptomatic infection before they end up in hospital as a breakthrough case so the more effictive the vaccine is against symptomatic infection the more protection the vaccinated will have against ending up in hospital.

The reason the protection against hospitalisation and sever disease is given is the vaccine gives further protection to a vaccinated person when a breakthough symptomatic infection happens in terms of their chance of ending up in hospital when compared to a symptomatic infection in a person of the same age who is unvaccinated.

This means the vaccine not only protects you against symtomatic infection it also protects you against hospitalisation and severe disease in the event of a breakthrough infection when compared to an unvaccinated person of the same age.

In professor Nolans model the effectiveness measures used by the HSE were are follows:

1. 80% protection against symptomatic infection
2. 93% protection against hospitalisation and severe disease

You will notice in professor Nolans model the vaccine reduction in the risk of hospitilisation is 65% this is the added benifit ontop of the protection against symptomatic infection the vaccine gives a vaccinated person. i.e. the 20% that have a breakthough infection (0.2) and the 35% that require hospitialisation (0.35) after the breakthrough infection (1-(0.2*0.35)) = 0.93 or 93%

In July the data out of Israel on the vaccine effectiveness against symtpmatic infection showed that it was waning down to 16% 6 months out from the 2nd dose with an average of 39% across the 6 months which was much lower than the 80% being assumed by the HSE in their modelling. However, the report out of Israel showed that the protection against hospitalisation and severe disease was holding at 91%

The HSE have only released the vaccine breakthough data in Ireland for the 12th and 19th of August and when I put this real data into their own model it indicated that not only was the protection against symptomatic infection much lower but the protection against hospitalisation and severe disease also had to be lower.

Since then, the initial report out of Israel to show the effectinvess of giving a booster dose of vaccine to boost the waning protection against symptomatic infection back up also reports boosting the waning protection against hospitalisation and severe disease back up by similar levels. This is the first time I had seen a report that confirmed what Nolans model showed with the real data interms of the protection against hospitalisation and severe disease being signifficantly lower than 93%

Just last Thursday a report on the analysis of the UK data also shows waning of the vaccines protection against hospitalisation and severe disease at the same rate (3% per month) as the protection against symptomatic infection.

https://www.youtube.com/watch?v=ERHPwt9ExL8&t=737s

The UK data also shows waning in the AstraZeneca vaccine

In the article you referenced Dr. Chun states:

"It's clear that the vaccines are highly effective, and without them we would be facing a much deadlier pandemic," Chun said. "As effective as the vaccines are, with emerging variants and increasing cases of breakthrough infections, we need to continue to be vigilant in taking measures such as indoor masking and social distancing."

He is correct, and even if the vaccine wanes down to 58% protection against hospitilisations which is the worst case scenario Nolans model indicated, with the real data, this is still highly effective. The issue is if policy decisions are being made on the assumption of 93% protection against hospitalisation and severe disease it could have serious concequences especially in Ireland where we have limited hospital capacity.

If we have a group of people where one of the group is infectious then the likelihood is that those who are unvaccinated may become infected but those who are vaccinated are very unlikely to become so ?

If the effectiveness of the vaccine to protect against symptomatic infection is 80% then a vaccinated person in the group is 80% less likely to get infected when compared to an unvaccinated person in the group. However, if due to waning the average protection is actually 39% then a vaccinated person in the group is 39% less likely to get infected when compated to an unvaccinated person in the group. Eitherway the vaccinated person is much less likely to get infected when compared to an unvaccinated person but the more the vaccine wanes the less protection they will have.

Unicorn Milk Latte

https://www.boards.ie/discussion/comment/117877426#Comment_117877426

It's also worth noting that in most cases 'breakthrough infection' still means minimal symptoms.

Severe cases occur for people which are immunocompromised, and have a serious underlying illness - for this group, boosters are ideal to improve protection.

Lumen

@actionmike wrote

vaccines are not 95% efficacious anymore

That is correct, but nobody has claimed that vaccines are 95% effective at preventing infection with the Delta variant.

Goldengirl

https://www.boards.ie/discussion/comment/117877426#Comment_117877426

Firstly the data you are inputting is data from other populations with different vaccine uptake , demographics and infection patterns than ours .

Secondly in assessing VE the age and comorbidities of people involved needs to be taken into account .

Mr.StRiPe

https://www.boards.ie/discussion/comment/117892122#Comment_117892122

Firstly the data you are inputting is data from other populations with different vaccine uptake , demographics and infection patterns than ours .

I only entered the real HSE data into the HSE model that demonstrates the effectiveness of vaccines so on the contrary I explicitly took into account Irelands vaccine update, demographics and infection pattern.

When I put the real HSE data into the model is showed that not only was the vaccine effectiveness against infection waning but the effectiveness against hopsitilasation and severe disease was also waning. When I posted this it was gererally accepted that the effectiveness against infection was waning 6 months out from receiving the 2nd dose the discussion was more on how much it wanes however, there was no evidence that the effectiveness against hospitialisation and severe disease was waning on the contrary it seemed to be holding firm at 90%+

The post where I referenced the latest report from Israel on vaccine boosters and the latest data from the Zoe group in the UK was to show that the data from other countries is showing the same pattern I got when I inputted the real HSE data into the HSE model. i.e. the vaccine effectiveness against hospitialisation and severe disease also wanes and at a similar rate as the waning against infection.

Secondly in assessing VE the age and comorbidities of people involved needs to be taken into account .

The age is taken into account as I am using the real HSE data in the HSE model for vaccine protection.

In terms of comorbidities you seem to be implying in Ireland that this disproportionately represents/explaines the breakthough cases admitted to hospital when compared to the unvaccinated cases admitted to hospital? can you provide a reference for this? as I have been trying to source this information myself.

Goldengirl

OK.

No reference except personal experience, oh, and every interview that has been given by any senior ICU doctor in the country, Dr C Mothefway for example.

Where are you from? You don't sound Irish if you don't mind me saying.

I am impressed though if you are for real.

Mr.StRiPe

https://www.boards.ie/discussion/comment/117890894#Comment_117890894

That is correct, but nobody has claimed that vaccines are 95% effective at preventing infection with the Delta variant.

The FDA claimed the Pfizer-BioNTech COVID-19 Vaccine is 91% effective with the Delta variant when they approved it on the 23rd of August!

https://www.fda.gov/news-events/press-announcements/fda-approves-first-covid-19-vaccine

Lumen

@Mr.StRiPe wrote

The FDA claimed the Pfizer-BioNTech COVID-19 Vaccine is 91% effective with the Delta variant when they approved it on the 23rd of August!

91 is not 95, disease is not infection, no mention of Delta.

Try harder.

Mr.StRiPe

https://www.boards.ie/discussion/comment/117893015#Comment_117893015

91 is not 95

You quoted another poster who said vaccines are not 95% efficacious anymore however, they did not mention in the quote, you referenced, if they were talking about being effective at preventing infection with the Delta variant or effective at preventing hospitialisation and severe disease. You concluded that they were referring to being effective at preventing infection with the Delta variant and the tone of your response indicated that nobody was claiming anywhere near this level of effectiveness at preventing infection with the Delta variant. I highlighted this is not true as the FDA claimed 91% effectiveness in preventing infection with the dalta variant when they approved the Pfizer-BioNTech COVID-19 Vaccine on the 23rd of August and provided you with the reference.

disease is not infection

You clearly didn't bother looking at the FDA reference I posted and the study the 91% effectiveness figure is making reference to. If you did you would have seen the effectiveness % the FDA are claiming the vaccine provides against infection is 91%.

The FDA claim the Pfizer-BioNTech COVID-19 Vaccine effectiveness against hospitialisation and severe disease is 95.3% or 100% based on the CDC definition of severe COVID-19

no mention of Delta

This is true however, there is also nothing to qualify the vaccine effectiveness %'s the FDA are claiming, as part of approving the Pfizer-BioNTech COVID-19 Vaccine on the 23rd of August, does not relate to delta which is the dominant strain of the virus in the US that the vaccine needs to protect against.

The FDA would be deliberately misleading the public to claim vaccine effectiveness %'s that are related to the alpha and/or beta variant and not explicitly qualify that was what they were referring to and it has been nearly a month since the 23rd of August and there has been no clarification from the FDA, as far as I'm aware, that they were not referring to the vaccines effectinvess against delta.

Surely you don't believe the FDA are deliberately misleading the public?

brickster69

https://www.boards.ie/discussion/comment/117897873#Comment_117897873

Some new figures from Pfizer

https://www.fda.gov/media/152161/download