Jean Fisch Profile picture
Analysis, rationalism & objectivity are my sins
Aug 10 6 tweets 2 min read
A little example to show well the assumptions of the Ioannidis "Lives Saved" study fit reality
- Take its IFR estimates
- Get the US MCOD covid deaths from CDC up to 14/11/20
- Calculate infx up to 31/10/20
Result? The study estimates that the US infx rate was 62% by 31/10/20😀 Image Now I took end October 2020 out of a reason: There are two sera studies for the US at that date
a) CDC: 8%
b) another study: 11% (side comment: actually 5% but then doubled for serareversion which is far too much)

thelancet.com/journals/lanam…

academic.oup.com/cid/article/74…
Jul 28 10 tweets 3 min read
I did the following:
- I took the sera studies considered as good by Ioannidis in his paper on IFR from 2023 for West Europe (incl DE/FR/SP/IT/PT/AT ...)
- I matched that with the info I have on covid deaths per certificate

Result? Covid IFR was >1% in Western Europe in 2020 Image Ioannidis published a piece early 2021 in which he selects 40 sera studies which he considers as good

I just went through them and focused on those that concern western Europe

This covers DE, FR, IT, SP, PT, AT, IR and NO so quite a lot!

1/ Image
Jul 26 8 tweets 2 min read
The paper by Ioannidis & al concluding "C19 vax saved ~1-4m" got published

As I flagged at the time of release⤵️, the paper relies on assumptions which are (often massively) too pessimistic vs the few real data points

But also its conclusions are marred by an error of logic

1/ The conclusion on lives saved depends on the severity of Omicron vs. Delta and Wuhan

By chance, Omicron proved much less intrinsically severe than Delta ... but we couldn't know this in 2021

So the study's conclusions on the legitimacy of measures are based on hindsight

2/
Jul 25 4 tweets 2 min read
Only the UK had a good idea of its infections and yet I am not aware of a single study using it to determine IFR of covid by age

A study was published today by Ioannidis et al on lives saved by vaccines and their IFR is massively lower than the real data based on from the UK

1/ Image Ioannidis et al rightly split the 70+ into community and care settings

I applied both to the population of England and Wales assuming 400k care places and hence 800k people of the 70+ (average stay = 2 years) exposed to covid via care settings

2/ Image
Apr 30 5 tweets 2 min read
This study looks well done but I don't see how it allows for "concerns about vaccines" as conclusion line

To take an analogy, if a study found higher car mortality in men than women wearing a seat belt, you would not expect "concerns about seat belts" as conclusion line

1/ I asked the authors about it also because, actually all studies of age-adjusted mortality by vax status show a higher rate among the unvaxed

So, without further further info about unvaxed, the conclusion to date here is that Moderna reduced mortality even more than Pfizer

2/
Apr 3 5 tweets 2 min read
If you are in covid twitter, you will see this headline pass your tweet today

"Quasi consensus in the French Medical academy for a lab origin of covid" relayed by the normally very serious @Sciences_Avenir periodical on science

Only issue? It is not true... Image @Sciences_Avenir What happened is that the academy
- got a presentation on a report of the covid origin, available here
- was asked to approve the conclusions of the report (see screenshot) which are on the measures to minimize lab risks NOT ON THE ORIGIN OF COVID ITSELF Image
Mar 16 7 tweets 2 min read
The Guardian claims that mortality rates have not yet returned to pre-pandemic trends in England + Wales

Here the latest view using standardized rates for death OCCURRENCES (not registrations)

As you see, mortality was essentially back on pre-pandemic trends by 2023 Image Now, if you follow mortality, you will probably say "hey, wasn't 2023 still showing a lot of excess?"

This is a spurious effect from death registrations. In 2022, the flu wave was huge and peaked in late December 2022, pushing deaths to only be registered in 2023

2/
Jan 28 6 tweets 2 min read
Statbel released its 2022 deaths by cause for Belgium (speed is not the essence)

Despite having asked to "report deaths even if covid only suspected", covid deaths Sciensano during the pandemic are 20% resp 40% lower than the actual deaths due to covid in Wallonia and Flanders Image Early in the pandemic, Belgium did something I still consider best in class

It instructed its medics to report any deaths which was only possibly covid, e.g. a person dies with covid in a home and suddenly 6 others die without test within days)

2/
Sep 7, 2024 16 tweets 4 min read
This tweet is a masterpiece of "writing to effect"

The statements on LC made by this Lab attracted much attention

Let go through them by adding the figs as they are known from the UK, the ONLY place in the world with a clue on its infx and potential LC

Hint: There is much spin STATEMENT 1: "incurable chronic disease"

This is technically correct: You can not take a drug and make it go away, there is only hope and prayers

BUT

1/
Jul 29, 2024 6 tweets 2 min read
Australia released its mortality data up to April 24

So far, 2024 is
- in line with the expectations from the Mortality Working Group when including pandemic years
- 5% above those from pre-pandemic trends (of which roughly half is directly attributable to covid)

1/ Image I still believe that a large part of the "unexplained" excess against pre-pandemic trends is actually just a "shift forward of the annual curves" from pink to blue (due to changes in immunity structures and covid being more contagious)

2/ Image
Jul 26, 2024 9 tweets 3 min read
Through its infection and Long Covid survey, ONS provides everything needed for ballpark estimates on the current situation for the 16-64

Risk of developing condition end 2023
Overall: ~1.5%
Severe: ~0.3%

% of pop with condition end 2023
Overall ~3%
Severe ~0.5%

Details below Image Everything comes out of the ONS winter infection survey

It is based on a random sample of the population which
/ got tested every week
/ self reported LC by
a) severity
b) when it started (less than 12 weeks ago, 12-53 weeks ago, etc.)

1/

Source: ons.gov.uk/peoplepopulati…
Image
May 29, 2024 14 tweets 5 min read
England and Germany have periodic randomized infection surveys which allows to estimate the average duration between infections

By inferring the missing data (more below), it looks as if the duration has remained roughly unchanged over the last year in England, at ~1.5 years Image The analysis is made more complex by the fact that the periodic infection surveys
- only started late in Germany
- was interrupted in England
(NL also has a survey but it is not randomized)

So one needs to "complete the data" to get an annual view and have a sense of the trend Image
Feb 18, 2024 18 tweets 7 min read
Here is the annual overview of weekly ASMR in western Europe

Netherlands is the only country which did not see significantly reduced mortality rates in 2023

and it is due to its 65+ and not its <65 and it should be a HUGE attention point for @rivm and @statisticscbs

THREAD⤵️ Image @rivm @statisticscbs Just to hammer home how much NLD is on another planet: Here is the annual growth rate of ASMR in Western Europe

Only one country goes up in 2023: Yep, NLD...

(I added a high value for the missing 2023w52 data to have full year for SWE and clearly also its ASMR will be down)

1/ Image
Jan 29, 2024 15 tweets 4 min read
NZE is the only country in the world with high vax, good death data and vax happening before and not in parallel with covid

It saw no excess in the<60, which is not attributable to covid

This de facto kills any claim of high vax deaths made elsewhere, let me explain⤵️ Image Just for clarity, I think it's vital to check for signals of covid vaccine induced deaths simply because there is risk eg
- intrinsic risk of novelty (of mRNA vaccines) even if the biology looks safe
- the appalling transparency record of pharma when things go wrong

1/
Jul 20, 2023 18 tweets 6 min read
I continue to see talks of high unexplained excess deaths in England and Wales in S1 2023

This is NOT confirmed by data by occurrence date and is an artefact resulting from
/ reporting by registration date
/ policy of only registering deaths once cause is assigned

THREAD Image Just to make my point crystal clear

The S1 2023 deaths by registration date excl covid are POSITIVE and high which signals lots of unexplained deaths

Yet, when using deaths by occurrence date (as one should), the same total is NEGATIVE. So in fact no/few unexplained deaths

1/ Image
Jun 1, 2023 7 tweets 3 min read
This paper making the rounds at the moment claims that "excess deaths started in Mar 2021 in GER and is unlikely to come from covid but matches vaccines"

It is rigged with issues which basically invalidate its claim

Let me go through the major ones⤵️

cureus.com/articles/14941… The key issue is that the authors did not take info account that in the absence of flu the winter peak will be displaced

Once you correct for this, excess deaths match covid deaths in 2021

1/

Jun 1, 2023 8 tweets 4 min read
Scotland has been reporting continues excess deaths (and this beyond covid) since November, even after corrections for mortality and population trend

However, the data by registration date is a bit misleading and a look at data by occurrence provides better insights ⤵️ Image Scotland provides data by occurrence by age on a monthly basis which allows one to make a proper assessement against expected deaths from population and mortality trends

And what comes through is that there are TWO periods of excess: "Xmas 22" and "Mar-Apr 23"

1/ Image
May 12, 2023 6 tweets 2 min read
This is actually quite insightful and shows how prior beliefs blurs and distorts the situation

Natural immunity was absolutely not divisive: It's value was clear but no gov was prepared to accept the costs associated with it until vaccines came

1/5 In particular, the impact of an "R0=2.5 / IFR >>0.5%" virus is such that no shielding mecanisms was deemed sufficiently secure to let circulation run high without taking uncalculated risks on the death toll

(here a little check I did a while back)

2/5

May 9, 2023 4 tweets 2 min read
Vaccine harm deserves better than the deplorable show that vaccine sceptics are offering so here an update of fatality from Germany (up to April 2022):

The death rate due to covid vaccination was at roughly 1.8 deaths per million doses administrated in early 2022 Image Quick add ons to forecome obvious questions

/ No need to remind me that, just as with long covid, fatality is not everything (I know): I could not find hospitalizations or sequels data

/ Unfortunately, the German monthly death data by cause is not split by age
Apr 8, 2023 18 tweets 4 min read
While Dr. Gupta's claim that new Danish data vindicated her low estimation of IFR in Mar '20 is incorrect, I still wondered how this estimate came about

This interview provides answers and highlights once more IFR's key role in a pandemic response ⤵️

From the interview, it is clear that Dr. Gupta’s low IFR estimate back in 2020 came from a set of considerations which reinforced each other

1) Covid-19 is a SARS virus and the world had been exposed to SARS before
-> Some immunity was there possible / likely

1/N
Apr 5, 2023 4 tweets 4 min read
In this @unherd column, @SunetraGupta asks the Q: "how wrong was I to suggest that average IFR was 0.05-0.10% back in 2020?"

The answer is
- "By a factor of 5-10 for DNK" (per official data)
- "By a factor 5-25 overall" if UK is anything to go by

It's quite simple really ⤵️ @unherd @SunetraGupta @DSTdk official 2020 death toll "FOR COVID" is 1,078

Source: dst.dk/en/Statistik/e…

@SSI_dk studies put covid prevalenced at 4.1% in Dec 2020

Source: pubmed.ncbi.nlm.nih.gov/34420152/

So the IFR in DNK in 2020 was 1,078/5,8m/0.04% = ~0,45%

ie "5-10x higher than 0.05-0.1%"

1/3 Image