Vaccine Adverse Event Reporting System
2021-04-03
VAERS, as it is commonly known, is the U.S. government's excuse for relieving the pharmaceuticals of all liability for their vaccines. Various people have estimated that anywhere from 1% to 10% of all adverse events are actually reported in the system. Participation is voluntary and open to all, and apparently there is no serious validation, so the data is not of high quality. For example, there are over 100 entries with completely bogus vaccination dates going all the way back to 1920. The CDC makes the data available here weekly, with a one-week lag.
The current drive to vaccinate everyone for COVID-19 is resulting in tens of thousands of entries in this database, and the analysis has turned up some interesting results. I have used SQLite Manager, a FireFox plugin, to perform some analysis. After the attached file is fed into SQLite Manager, the CDC data can be easily imported. The included views turn up some striking factoids.
The VAERS_ID columns are presumably automatically generated by the database, such that each new case has an ID which is one higher than the last case. Scanning the (first) ID column in the DATA table, it is obvious that there are gaps in the numbering, leading one to wonder if some of the data has been removed. More astonishing, sometimes there are huge leaps in the numbering for a run of cases, followed by resumption of the expected numbering. All in all, as the AWOL view shows, over 80% of the entries seem to be missing. A suspicious person might wonder what happened to them.
A look at the cases by age range (covid_by_decade) shows that, although there are plenty of cases for the under-60 crowd, there are, unsurprisingly, extremely few deaths. This same view shows that women of childbearing age (20-50) are four times more likely to report an adverse event than men. For all age ranges women are more at risk, though some of this might possibly be because women are perhaps more likely to report untoward symptoms than men.
A breakdown by state (covid_by_state) shows that, for whatever reasons, some states report up to five times more adverse events per capita than others.
Breaking the cases down by manufacturer (covid_manufacturer_stats) shows that about half are from Moderna, and half from Pfizer\Biontech. Eyeballing the complete breakdown by lots (lot_counts), we see that most lots are relatively harmless. This is confirmed by the safe_lots view which shows that at least 93% of the lots have produced only 7% of the reported cases. Many of the lots have only one case, so it seems likely that many lots produce no cases at all, and are therefore not in the database. Thus the number of lots that produce very few, or no, cases is probably over 95%. The flip side of this is that less than 5% of the lots produce 93% of the cases. Clearly not all lots are created equal, as depicted by the dangerous_lots view. Each of these 19 lots produced over 100 cases, over 10% of them fatal.
