Having had my second mRNA booster with the bivalent vaccine, I noted that another virus was taking over sections of the internet. The idea that the more vaccines you have, the greater your risk of COVID19. In Mercola land and elsewhere, this has now become established fact and adds to calls for public health officials to be prosecuted as criminals and generating copious outrage and hate.
The paper immediately struck me as worth exploring. The estimation was made that vaccine effectiveness at preventing infection was 30%; that is encouraging, even if not in line with other research, but headlines were made by the claim that the more vaccines you have, the greater your risk of COVID infection. The paper comes with the significant warning that it has not been peer reviewed. Smells fishy to me, so let’s have a look.
The Summary of the paper is a follows:
“Among 51011 working-aged Cleveland Clinic employees, the bivalent COVID-19 vaccine booster was 30% effective in preventing infection, during the time when the virus strains dominant in the community were represented in the vaccine”.
For one thing, preventing any infection seemed odd, as we know that the vaccines available now primarily prevent severe illness: Last winter in the USA monitoring showed:
“From January to April 2022, when the Omicron variant was predominant, hospitalisation rates were 10.5 times higher in unvaccinated persons and 2.5 times higher in vaccinated persons with no booster dose, respectively, compared with those who had received a booster dose.“
A closer look at the paper reveals some big errors.
The graph in which they display the outcomes for the different groups looks impressive, but in fact the vertical axis on the left shows that the real differences are quite small. Nevertheless, how real are they?
A glaring problem, the wobbly foundation of this research, is that they separated the never infected from the previously infected on the basis of self reported positive PCR tests. This misses those with asymptomatic infections (about a third of cases) and those who just didn’t want to do the test. So the number of people who they called never infected was, simply put, wrong.
The prevalence, according to the CDC in that area at that time of antibodies to SARSCov-2 was about 60% as opposed to their 40%. Indeed, given that the study was one of health workers, the real rate would be very likely to be much higher. So a big chunk of their non-immune group would have had prior immunity.
That represents a hefty does of selection bias. There is certain to be big differences between the two groups which will not be accounted for by their statistical methods. Hospital workers who had not been vaccinated by 2022 would include those who don’t trust the vaccine, think its dangerous, believe in conspiracy theories, or think COVID19 is just a little cold,
They will also be less likely to have self tested for cold like symptoms and so will be in the ‘uninfected’ cohort in error. Indeed, even if they are sick they will be less likely to self test as a positive result would,(indeed should, in the case of healthcare workers) be associated with at least a little embarrassment.
Not comparing groups accurately really does mess with their conclusions and renders their findings, as I said, fishy. Indeed, the groups may also be influenced by face to face healthcare workers being far more likely to be vaccinated and also more likely to get infected. Who knows?
The the global headlines came from a single sentence within the paper. This is due to the fact that it is only arrived at by poorly defined complex statistics. That is a little naive on the part of the authors who might have know such a statement would be seized upon, amplified and used to create sensation.
That the paper hasn’t been peer reviewed in also important here. Currently it is listed on MedArchive and has not been critically analysed. I really do think that when it is, it will be changed significantly and improved before publication. Indeed, peer reviewed journals might reject the paper for the reasons I list above. In normal circumstances that would have been the end of it, but instead of that it has had 32,000 views on Twitter alone and is being used by antivaccine folk like Mercola and his many acolytes and sadly by the avuncular bur deceptively inaccurate Dr John Campbell and others.
Pre-prints are in many ways unedited and should not be used for decision making.
Thus the spread of the uncritical notion that vaccines cause infection has become fixed in many peoples minds despite the best effort of bloggers like the hilarious Culture Catz who is also on the ball with science, but unfortunately has far less reach than the outrage machines. Likewise the excellent Debunk the Funk.
In fact, peer reviewed studies around that time showed that unvaccinated had 10 times the risk of hospitalisation than the vaccinated. Even without a booster the vaccination reduced the risk by a factor of 2.5. This study was peer reviewed and methodologically sound.
Vaccination is changing all the time. This is fine – the choice is yours, but mistakes are made if decisions are based in low quality evidence like this. The misplaced anger generated by these claims and currently Woozling along in cyberspace is not good for health either.