Digging Deeper Into EudraVigilance (part 3)
A word about record duplication
Death is bad enough without it being duplicated and in part 1 of this series I highlight the problem we face if we try and count fatalities using the information provided for the public on the EudraVigilance website. Fatalities are classified by reaction group and if somebody dies because of cardiac arrest and only cardiac arrest then there isn’t a problem. However, if other things go wrong then that death will be counted under each and every pertinent reaction group. This problem is overcome by using the line report system and downloading the raw data, with each record representing one individual. Except that there is another duplication wrinkle…
Individuals will be represented twice in the database if they switch between products. Thus somebody whose first dose was AstraZeneca’s Vaxzevria who is then advised to switch to Pfizer’s Comirnaty will have a record under each product. I’ve gone through the raw data looking for these switches and offer this colourful screenshot so you get a feel:
What we have to do here is de-duplicate the record system whilst at the same time retaining information pertaining to all vaccines received. Incredibly, some people received a cocktail of three different vaccines prior to death. It took half a tray of blueberry muffins and four pots of tea to clean, re-code and de-duplicate the data such that it will be of use in statistical analysis.
I can report that after triple-checking spreadsheets EudraVigilance yields 23,891 unique fatalities from 20 Dec 2020 to 28 Feb 2022. A total of 614 records were found to be duplicated as a result of switching vaccines and the following table summarises the situation:
Subscribers may well be scratching their heads over the 2,092 zero suspected vaccines. These are instances where a vaccine has been given but the healthcare professional has decided not to register them in the suspected drug field. No matter, for we can remedy the situation by crosstabulating with manufacturing source data:
This is a mighty interesting table in its own right, with AstraZeneca leading the triplicates and duplicates as folk were quietly discouraged from taking second and third shots. In contrast, Pfizer leads the way in zero suspected shots. Do I smell product loyalty among European healthcare professionals1?
It would seem that not all countries were created equal when it comes to the rigour of healthcare reporting. However, the sun is over the yardarm and my stomach is rumbling. Time to put some toast under the grill, methinks.
Kettle on!
Germany leads the way with 23% of cases recorded for Pfizer products, with the Netherlands coming second at 16% of cases recorded, and France getting the bronze with 14% of cases, these three nations alone accounting for 53% of all reported events.