Crawford on Vaccine-Induced Deaths and Ignored Safety Signals

Data analysis reveals two issues that are going to be very important over the next few years. What level of vaccine-induced deaths are hidden in Covid death statistics? And, why has the CDC deliberately chosen metrics that hide any signals that a vaccine may have safety problems?

The Covid panic, like other world events, has shown again that there is great talent outside the standard institutional framework. One math-data-stat guy has proven to be a real treasure over the past two months.

Mathew Crawford blogs at Rounding the Earth on Substack. Substack is a remarkable platform for quality writers who may have been silenced, sacked, or simply contrarian. People like Alex Berenson, Spartacus from Catallaxy, Bari Weiss, Matt Taibbi and others.

This is an introduction to two of Crawford’s series of posts which are both incredibly important. They illuminate very serious issues in Covid vaccine world which will haunt our political and public health leaders for years to come.

Crawford has made such a splash that a Twitter moneybags offered a $1m reward to counter his arguments.

At 360 million doses delivered, these estimates suggest between 72,000 and 180,000 (or maybe even a little more) vaccine-induced deaths in the U.S. during the experimental COVID-19 vaccination program

Mathew Crawford,
Estimating Vaccine-Induced Mortality 1

Vaccine-Induced Deaths Hidden in the Data

Introduction to the posts below to help non-math readers: In August 2021 Crawford started sniffing around the US Vaccine Adverse Events Register (VAERS). He had recognised that much of the damage from the vaccine follows a similar mechanism to the one used by SARS-Cov-2. He started crunching numbers when he learned that many vaccine-induced deaths were probably being recorded as Covid-19 deaths in the VAERS database (the McLachlan paper). Also worrying was that the CDC wasn’t investigating reported deaths, and that autopsies were being stopped.

He created daily Case Fatality Rate (CFR) numbers to show any dramatic changes after vaccination programs began. Reported CFRs are just big totals, not day-by-day. If daily CFR went up unexpectedly, it would indicate vaccines were doing damage, because Covid lethality wouldn’t change overnight.

He aimed to match deaths to when their cases were diagnosed. Because people who die from Covid on average die 18 days after symptoms appear, he created daily CFRs based on deaths for a particular day, compared with case numbers from 18 days previously (the lag he mentions in the posts). To make up for lumpy case reporting he used a running 7 day average (the smoothing he mentions).

He then checked the 30 days following the beginning of vaccination programs across the world. He found a large spike in daily CFR after vaccination in about 85% of nations. If the CFR didn’t change, often both cases and deaths rose dramatically.

He then looked at European data and calculated CFR in the same way. To pull all that data together in a comparable way, for each country he set the CFR at the beginning of the vaccination program to an index of 1 (the normalising he mentions). That way any increase in CFR after vaccination turns up as a multiple of the CFR on Day 0 of the vaccination period in each country.

The excess CFR (CFR > the normalised 1) that appears after vaccination is an estimate of the people dying from vaccine injury who are hidden in Covid-19 death statistics. For example, a normalised daily CFR of 2 would suggest about half of the people who died from Covid on that day died from vaccine-injury.

Crawford refined his estimates in later posts, listed below.

Read the lot. They matter.

This means that the experimental COVID-19 vaccination programs may be killing somewhere between 200 people per million doses and 500 people per million doses—perhaps even more since the U.S. has a more substantial population living with substantial comorbidities, and the world’s best cardiac trauma care. At 360 million doses delivered, these estimates suggest between 72,000 and 180,000 (or maybe even a little more) vaccine-induced deaths in the U.S. during the experimental COVID-19 vaccination program. As we will see in future articles, this estimate range matches numerous other mortality signals.


Vaccine Safety Signals Squashed by Poor Design

Health and regulatory authorities have an obligation to monitor drug and vaccine use for signals that indicate that they may be, or are becoming, unsafe.

Crawford highlights that the definition of the key metric the CDC uses as a safety signal (the Proportional Reporting Ratio, PRR) is scale-invariant. This means it doesn’t matter how many adverse events there are, so long as the proportion of particular adverse events stays the same. All relevant references to CDC are in Crawford’s posts.

The PRR is a comparison of two ratios. It finds the proportion a single type of adverse event against all adverse events for a drug/vaccine, then compares it to the same ratio for a comparable drug/vaccine.

For example, (made up numbers) if 0.5% of all adverse events for Comirnaty are myocarditis, and 0.5% of all adverse events for a comparable vaccine are myocarditis, then there is no signal. The adverse event proportions are the same.

The problem is when you compare proportions, the absolute numbers don’t matter. You could have tens of thousands of cases in one drug/vaccine, and hundreds in the other, but if the proportions of adverse events are the same you wouldn’t notice. Check Crawford’s explanatory example:

This is the basic issue. Crawford has had substantial to-and-fro on it. Again, read all the posts. The math isn’t too bad (though you could skip the Chi-Square debate).

The implication is simple. A drug/vaccine that has a huge number of total adverse events (Covid vaccine adverse events make up a large percentage of VAERS records) can have a huge number of a specific adverse event and not trigger the safety signal.


Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.