Analyzing Japan VAERS death reports

Over 100 deaths were considered to be related to the mRNA vaccine by actual health care professionals

Key takeaways:

• Japan is probably the only country where healthcare professionals are willing to go on record in the VAERS writeup that a death was actually related to the mRNA vaccine

• There are three levels of certainty within the reports - related, probably related and possibly related

• There are 169 deaths related to the vaccine, and 158 of these reports were by healthcare professionals

• There are 20 deaths probably related to the vaccine, and 14 of them were by healthcare professionals

• There are 186 deaths possibly related to the vaccine, and 154 of them were by healthcare professionals

• There are 965 deaths related to cardiac issues, and 795 were made by healthcare professionals

• I hope the CDC does not stop publishing VAERS writeups for Japan

• Click here to see the full dataset

I have not done any analysis of the deletions of foreign reports till now. But recently, I noticed that the largest number of deletions (followup reports) in the foreign dataset was from Japan¹.

So I started looking into the Japanese VAERS reports, and accidentally discovered something very interesting.

It is probably the only country where the Doctors are willing to go on record in the VAERS report stating that the death was related to the vaccine².

So I parsed the writeups to look for text patterns which suggest that the death was related to the vaccine.

If you are interested, this is what the code looks like:

169 death reports were considered related to the vaccine, and 158 were made by healthcare professionals

If you filter for IS_RELATED = ‘Yes’, there are a total of 169 reports.

If you filter further for IS_HCP=’Yes’, there are a total of 158 reports.

Probable, plausible, possible

I want to take a quick detour to discuss the usual meanings of the words probable, plausible and possible, as it applies to probability of events.

They usually indicate varying levels of certainty, and this picture explains it well

Source

As you can see, probable means highest level of certainty, plausible is less so, and possible is the lowest level of certainty.

In addition to the reports which plainly state that the death was related to the mRNA vaccine, there are many death reports which claim that the causal association is “probable”, “possible” or “cannot be excluded”

I group “possible” and “cannot be excluded” under “Possible” in my analysis.

20 death reports probably related to the vaccine, 14 made by healthcare professionals

If you filter for IS_RELATED = ‘Probable’ you will see 20 reports.

If you also filter for IS_HCP=’Yes’, you will see 14 such reports.

186 death reports possibly related to the vaccine, 154 made by healthcare professionals

As I mentioned before, I grouped the reports which said “possibly” related to the vaccine as well as “cannot be excluded” into a single IS_RELATED category called “Possible”.

There are a total of 186 such death reports, and 154 of them have been made by healthcare professionals.

965 cardiac related deaths, 795 made by healthcare professionals

I filtered the ENTITY_LIST field by the word ‘cardi’ which is a reasonable proxy³ term for cardiac related issues. I generate the ENTITY_LIST by using a generic biomedical NER model called scispacy to extract all the medical jargon in the writeup, as described in the article below:

You can the filter for IS_HCP is true and see that 795 of them were made by healthcare professionals.

Is the medical community normalizing myocarditis?

Link to tweet

1. These are not specific to COVID19 vaccines, but cover all reports for all vaccines. I plan to analyze these deletions for a future article, but I will not be writing about VAERS deletions in this article.

2. Now that this is public knowledge, I wonder if the CDC is going to stop publishing the writeups for Japan next?

3. It will probably have some false positives but they are very rare