How many participants in v-safe had serious adverse events?

Using the Most Severe Outcome to properly categorize adverse events

Summary:

• This is the first analysis (that I am aware of) which properly categorizes the severity of outcomes in v-safe

  • Using the “Most Severe Outcome” for each registrant boils down the analysis to one unique value per registrant and makes it much easier to discuss v-safe outcomes

• The v-safe statistic that is usually cited by Dr Peter McCullough and ICAN folks is incorrect

  • However, the v-safe statistic cited by CDC, and parroted by the vaccine pushers, is even worse - it is wrong by an order of magnitude!

• ICAN sued the CDC to publish the v-safe free text entries (they initially refused, claiming the MedDRA codes were sufficient). This analysis also directly refutes Dr Dan “Debunk the Funk” Wilson’s video explaining why the ICAN lawsuit is a “scam”

  • Most v-safe events do not occur within the first 7 days, so limiting the analysis to the first week makes no sense at all

  • The “Other” category for Most Severe Outcome can only be properly analyzed by performing text analytics on the v-safe free text entries, while Dan Wilson thinks these entries are not important

Dr Peter McCullough seems to be having a hard time keeping his v-safe numbers straight.

Remember, these were the five choices provided in the v-safe app if someone checked the “Get care from a doctor or other healthcare professional” box.

Source

Suppose we assign them levels based on how severe these outcomes were

1 : Other - please describe¹

2 : Telehealth, virtual health, or email health consultation

3 : Outpatient clinic or urgent care clinic visit

4 : Emergency room or emergency department visit

5 : Hospitalization

Peter McCullough tweeted this in Feb 2024:

Link to tweet

I have added the corresponding labels in brackets:

V-safe: 7.7% of COVID-19 vaccinated (n=10,108,273) acutely so sick with side effects they go to clinics [2], urgent care [3], emergency room [4], or are hospitalized [5]. NOT "rare."

Then he tweeted this in April 2024:

Link to tweet

7.7% vaccine recipients get so sick they land in urgent care [3], emergency rooms [4], or become hospitalized [5]. No wonder there is tremendous vaccine hesitancy!

In a recent interview in August 2024 he said:

“What we know is From the CDC v-safe data, 7.7% of Americans who took a shot got so sick they had to go to the emergency room [4] or be hospitalized [5]. 7.7%. That's a sample size of 10 million people. That's how toxic these shots were.“ (at 24:33)

Source

First of all, it does bother me that the number 7.7% did not change, but the list of included items kept getting shorter (indirectly implying more severe outcomes on average for the 7.7%).

Second, since none of these include the category called “Other”, all these numbers are incorrect².

So let us define the Most Severe Outcome (MSO) for each registrant as the outcome with the highest level in my chart, with 1 being the lowest severity (Other) and 5 being the highest severity (Hospitalized).

Since a registrant can only have one MSO, it substantially reduces the complexity of the calculation and also improves the presentation of the data.

For the rest of the analysis, I will be using the Vaccine Enthusiast Pfizer Cohort which I described in my previous article.

As you will see, even though the vaccine skeptics like Peter McCullough and ICAN did not get their numbers entirely correct, they were still a lot more correct than the CDC!

Calculating the MSO counts

I wrote a Python script to identify the Most Severe Outcome for each registrant in the Vaccine Enthusiast Pfizer Cohort who reported seeking medical care (about 83K people).

Remember that the total number in the cohort is about 1.3 million people. 

Click here to access the full dataset

The algorithm is pretty simple and looks like this:

most_severe_outcome = 'Other' # default value
if checkin mentions consultation:
  most_severe_outcome = 'Consultation'
if checkin mentions clinic visit:
  most_severe_outcome = 'Clinic'
if checkin mentions ER visit:
  most_severe_outcome = 'ER'
if checkin mentions hospitalization:
  most_severe_outcome = 'Hospital'

Let us look at a breakdown of the Most Severe Outcomes. The percentages below are out of 83K and not out of the original 10 million.

Link

1 : Other - please describe = 14.6%

2 : Telehealth, virtual health, or email health consultation = 20.7%

3 : Outpatient clinic or urgent care clinic visit = 43.3%

4 : Emergency room or emergency department visit = 11.8%

5 : Hospitalization = 9.6%

Of the ~83K people, 12124 people, which is more than 14%, had Other as the Most Severe Outcome. In other words, people who reported getting medical care do not mention any of the 4 checkbox choices actually constitutes a big chunk of the total number of people who reported seeking medical care.

Within the original list of 1.29 million registrants, this makes up 12124/1.29 million = 0.93%.

When you calculate the rates for the checkbox selections (out of 1.29 million registrants), these are the numbers you get:

Total: 83201 = 6.45% of 1.29 million reported medical care.

Of these,

2 : Telehealth, virtual health, or email health consultation = 20.7% = 1.34%

3 : Outpatient clinic or urgent care clinic visit = 43.3% = 2.79%

4 : Emergency room or emergency department visit = 11.8% = 0.76%

5 : Hospitalization = 9.6% = 0.62%

[Sanity check: 0.93% + 1.34% + 2.79% + 0.76% + 0.62% = 6.45%]

Now contrast this with what the CDC reported in their Lancet paper for Dose 2 Pfizer:

As I mentioned in my previous article, the huge discrepancy is due to the fact that the CDC considers a 7 day window in its analysis, and only considers Doses 1 and 2 and ignores reports for Dose 3 and later.

What percentage had severe adverse events?

Let us take another look at the categories:

1 : Other - please describe

2 : Telehealth, virtual health, or email health consultation

3 : Outpatient clinic or urgent care clinic visit

4 : Emergency room or emergency department visit

5 : Hospitalization

Suppose you consider only categories 3, 4 and 5 as SEVERE adverse events, this means 2.79% + 0.76% + 0.62% = 4.2% of registrants reported an event which could have been a severe event.

As I mentioned in a previous article, this number is much closer to the one mentioned by Peter McCullough, Del Bigtree etc, and an order of magnitude higher than the number implied by the CDC v-safe paper.

Comparison with the full dataset

We can run SQL queries over the full Consolidated Health Checkin CSV file as a sanity check.

I created a second table from the original CSV file (which has 146 million rows) for all the checkins where the user selects one of the Health Impacts as “Get care from a doctor or other healthcare professional“

There are about 1.06 million rows.

If you look at the number of people, you get 797396, which closely matches the 782K number discussed by Del Bigtree and the team at ICAN in their videos.

Of these, 71911 has Hospitalization as the Most Severe Outcome

And 107112 had ER visit as their most severe outcome:

And 336947 had clinic visit as their most severe outcome:

And 163307 had consultation (e.g. telehealth) as their most severe outcome:

This leaves out 118119 with an outcome of Other.

Percentages out of 797396:

1 Other = 118119/797396 = 14.8% (14.6% in the PVEC cohort)

2 Consultation = 163307/797396 = 20.4% (20.7% in the PVEC cohort)

3 Clinic = 336947/797396 = 42.2% (43.3% in the PVEC cohort)

4 ER Visit = 107112/797396 = 13.4% (11.8% in the PVEC cohort)

5 Hospital = 71911/797396 = 9% (9.6% in the PVEC cohort)

As you can see, the percentages are quite similar and suggests we are on the right track.

I did not run the Python script on the full dataset as it is

a) very time consuming

b) I do not have enough space to upload it to my Zoho Analytics account

c) I do not have enough space on my GitHub Pages repo to upload the JSON files³

If anyone wants to reproduce this work on the full dataset, please get in touch and I will send you the Python script.

What else we can infer from this analysis

This analysis also disproves two more claims made by Dan “Debunk the Funk” Wilson.

He says this:

So if you look at the V-safe data, the greatest number of reported events occurred on day one following the dose of vaccine being given.

[00:11:01]

From there, the average number of reported events declines each day. So even if they're just looking at the first seven days, you're capturing most of the reactogenicity that is happening with these vaccines.

Link

But the data says otherwise.

For example, in my analysis, the number of people who were hospitalized within 7 days of their most recent vaccine dose was only 425.

While the number hospitalized within 42 days (considered standard for most vaccine safety analysis) was 2197, nearly 5 times the number!

Dr Dan Wilson also says the following (emphasis mine):

[Dr. Wilson]

So that's just him explaining the free text fields that I mentioned earlier. That was the whole point of this current lawsuit. And he doesn't explain what they expect to learn from it or why they are so important. Because really they're not. Unless you do some epidemiological statistics on them, which you're probably not going to be able to do with that kind of data, you're not going to learn anything.

[00:07:05]

This is just totally wrong, and in fact we can prove it using just the current topic of discussion.

As I pointed out before, there are 12124 registrants whose Most Severe Outcome was “Other”

So what kind of medical care did these people get? Was it something serious?

The only way to know, is to read the free text boxes and do some kind of text analytics.

How is it possible for anyone, let alone an actual Doctor, to casually dismiss all this information? This is why I keep pointing out that the current vaccine safety review process is third-rate.

1. This gets assigned the lowest level of 1 because we do not have any way to know how severe it was. When you read the free text entries, you can see that most of them are non-severe and in fact many of them are not even necessary, such as someone mentioning that they went to the Doctor to get the recent vaccine dose

2. Just to be very clear, I am not blaming Dr Peter McCullough. I still quite strongly believe that this data analysis should have been an open source project with plenty of input from the Data Science and Natural Language Processing community. Big Tech’s censorship of the scale and severity of COVID19 vaccine injuries ruined the possibility of a timely discussion of COVID19 vaccine injuries, because most people who had the skills to help out had no idea what was going on.

3. These are the files you use for the v-safe timeline visualization tool