The best research paper on mRNA vaccine safety

The Qatar research paper cited by Kevin Bass is a well conducted study. But does it actually show that the mRNA vaccine is safe?

Key takeaways:

• The Qatar study published in early 2023 is probably the best population wide study of the mRNA vaccine safety, but it does not make a convincing case for the safety of mRNA technology

• A 1-in-100K¹ (people) death rate is too high in my opinion. Interestingly, the actual rate is between 1-in-100K and 1-in-1M doses, but the final rate per person goes below 1-in-100K because of a multi-dose schedule. In that case, why choose a vaccine with multi-dose regimen?

• The causality algorithm used to determine if the vaccine caused the death is too stringent

  • The risk factor of heart disease makes it “low probability of causation”. Then why were these high-risk groups given STRONGER recommendations to get vaccinated?

  • The days-cutoff of 5 days to assign “high probability” is arbitrarily small

• The detailed list of anonymized patient information is missing important information

  • Days to symptom onset

  • Detailed reports not provided for all the deaths. Which ones were left out, and why?

In a previous article I wrote this:

My first objection is just philosophical. I don’t think vaccines should be so dangerous that they show up in aggregate mortality rate calculations.

Source

And I wrote that because it was quite easy to see the problem with this approach using just plain math.

I recently came across a paper which can help me explain my view.

Here is what Kevin Bass wrote on Twitter recently:

Link to tweet

There was a followup question from OpenVAERS asking for citations, and he cited the Qatar study.

The Qatar study is easily one of the better studies on this topic. In fact, they provide a supplementary table with an example list of deaths which they considered during the study. They even report the age decile in this list.

But does the paper actually make the case that the mRNA vaccine is safe?

Let us look at it in some detail.

Why is a 1-in-100K death rate considered OK?

Even this very careful study, which used a pretty high bar to associate the vaccine to a death (discussed later), found a rate of nearly 1 in 100,000 deaths.

while death rate among the vaccinated persons with either high or intermediate probability of relationship to SARS-CoV-2 vaccination was 0.98 per 100,000 (8 deaths classified as high probability and 15 deaths as intermediate probability among 2,347,635 unique persons who received at least one dose of a vaccine)

To his credit, at least Kevin points out that 1-in-100K is “highish”.

This number means there will be 3300 deaths in the US alone (US population = 330 million).

And it translates to 78000 people if you consider the whole planet’s population.

How is that OK for a vaccine?

Would people be OK if that happened for a traditional vaccine given to children?

We would definitely hear about it in the news, isn’t it²?

The causality algorithm is too stringent

I am not sure what is going on with the medical community nowadays, because we have now moved from “blame the vaccine until you are sure it did not cause the bad health outcome” to “don’t blame the vaccine until you are sure it did cause the bad health outcome”.

This low bar is already bad, but it is much worse if

a) the vaccine companies sell to the government instead of selling directly to consumers

b) governments then create quasi-mandates for vaccines by threatening people’s livelihoods

So here is the algorithm used to determine causality:

Notice that the “High Probability of relationship to vaccine” will only be true if three conditions are true:

1) No clear alternate cause of death

2) No risk factors that could reasonably have contributed to death

3) Death within 5 days of any vaccine dose administration

As you can imagine, this is actually a pretty high bar to meet especially if you add conditions like “no risk factors that could reasonably have contributed to death”. Why is that? Because existing heart disease is considered a major risk factor which is supposed to lower the probability that the death was due to the vaccine.

To make this point more clear, I scraped the table information from the PDF file and created another Zoho Analytics dataset.

Click here to see the dataset

The risk factor is very common

Let us filter the dataset for ‘Contributory factors’ contains the keyword ‘heart’ (usually indicating prior heart issues)

Notice that all of them are automatically intermediate or low probability, including a death which happened after just 6 days.

There is also a second problem with this approach - heart disease is already a leading cause of death in Qatar (like it is elsewhere), so using this risk factor automatically lowers the probability that this was due to the vaccine.

Source

Just to be clear, I am not suggesting patient history should be ignored. I am simply pointing out that it lowers the bar for the vaccine³.

The date cutoff of 5 days is arbitrary

Next, let us consider the arbitrary cutoff of 5 days. Why should we stop at 5?

If we double this number to 10 (why not?), and decide that every death within 10 days is more likely to be due to the vaccine, the mortality rate will go up.

Right now, they estimate 23 deaths (8 high probability and 15 intermediate probability) to the vaccine.

Due to Qatar’s low population, the mortality rate calculation is so sensitive that even an increase of 2 more people added into the “high or intermediate probability” bucket increases the mortality rate by 8%, and 3 more people added increases it by 13%! (23 deaths vs 25 deaths vs 26 deaths for 2.3 million doses)⁴.

Missing information

Even this otherwise very thorough study is still missing some important information.

Days to symptom onset is not provided

If you look at the table, you will notice that the days to symptom onset is not provided.

Instead they use the Days to Death in the calculation.

This is quite problematic.

This means if the patient noticed a symptom immediately after vaccination, but died more than a month later, they would not even be a part of this study.

This is why I keep insisting that VAERS, despite all its flaws, is still the best vaccine injury database. Just like Australia, the Qatar vaccine injury dataset (to the extent that you can call this paper a dataset) omits even something as basic as “days to symptom onset”.

Detailed reports are not provided for all deaths within 30 days

There were a total of 138 deaths within 30 days of vaccination.

Among 6,928,359 doses administered, 138 deaths occurred within 30 days of vaccination; eight had a high probability (1.15/1,000,000 doses), 15 had intermediate probability (2.38/1,000,000 doses), and 112 had low probability or no association with vaccination.

But the supplementary table lists detailed reports for only 52.

What if all the other 86 deaths were in the younger cohorts?

What if all the other 86 deaths had no contributory factors?

Why are we asked to take the word of the authors that they have done a thorough analysis, and have followed their algorithm for every single patient?

Reverse engineering an acceptable vaccine death rate using crude death rates

(Update: please do read the comment by Aletheia after reading this section)

The paper also mentions the following:

Crude death rates in Qatar for the years 2019, 2020 and 2021 were 6.60, 7.94, and 8.74 per 100,000 population

In other words, the crude death rate changed from 6.6 to 7.94 from 2019 to 2020 - this is a difference of 1.34 per 100K. Let us say all of it is attributable to COVID19. So COVID19 was responsible for an excess death rate of 1.34 per 100K.

Ideally, you would want a vaccine with no more than a 10th of that death rate⁵. So we would expect about .134 deaths per 100K, which means no more than 3.16 deaths directly attributable to the vaccine (based on ~ 2.3 million doses administered). But we already know that 8 deaths were attributed to the vaccine with “high probability”.

On top of that, we would add these other considerations⁶:

a) the jump from 6.6 to 7.94 was probably also due to missed healthcare visits and not all directly due to COVID19, so the 1.34-per-100K number should probably be lower

b) if you make the causality attribution less stringent, that would add more people into the “high probability” bucket, making the actual vaccine caused deaths even higher

c) at the limit, if all the 138 deaths within 30 days of vaccination were caused by the vaccine, that would be a death rate of 6 per 100K, almost matching the crude death rate for 2019!

So 8 deaths over 2.3 million doses equals about 0.34 deaths per 100K doses.

And 138 deaths over 2.3 million doses equals 6 deaths per 100K doses.

If you use my oversimplified metric, then I can in fact argue that at best the vaccine is 0.34/1.34 = 25% as bad as COVID19, and in the worst case 6/1.34 = more than 4 times as bad as COVID19!

So even if we “compare” COVID19 and the vaccine, we are not able to definitively say that the vaccine was extremely safe.

Summary

While this might be the best study on mRNA vaccine safety due to the excellent safety monitoring in Qatar, in my opinion it does not really make a convincing case that the vaccines are “extremely” safe even if you accept the mortality rate calculated by the authors.

1.

2. At least we would have heard about it if it had happened a decade back. I am concerned we are now beginning to normalize post-vaccine deaths, which is good for no one except the Pharma companies.

3. It is also quite ironic, because people with ‘existing heart conditions’ is the same cohort which is considered vulnerable, meaning they got a stronger recommendation to get vaccinated.

4. Of course this also works in the other direction - meaning if we reduce the cutoff days to 3, it will reduce the mortality rate quite a bit.

5. Well, ideally you want it to be zero! But for some reason vaccine promoters always want us to compare the vaccine with the disease itself. My point is that even this comparison is not looking good for the COVID19 vaccines.

6. I would not be surprised if I have oversimplified this paragraph and maybe missed some important factor which substantially changes my analysis. If you can think of any please do let me know in the comments.