This article was originally posted in the
Recently, some people have been spreading the idea that getting additional doses of the COVID vaccine increases the risk of catching the virus. The suggestion was made in an and by Florida Governor Ron DeSantis. The notion seems to stem from last December by researchers from the Cleveland Clinic. Opponents of vaccines have been using it to argue their case, worrying a fair number of people, if the emails I have received on the subject are any indication.
However, it would seem as if most people have not actually read the paper in its entirety, because what it actually found was that the bivalent vaccine reduced the risk of getting infected with COVID-19.
First, it鈥檚 worth explaining what we mean by a preprint. Most scientific research is presented at scientific meetings and published in scientific journals. Preprints are scientific manuscripts uploaded to the internet without any external peer review. The potential for problems was illustrated back in 2021 when a University of Ottawa study suggested that the rates of post-vaccine myocarditis were several-fold higher than previously reported. That The researchers had miscalculated the number of vaccine doses administered in Ottawa during that time. They eventually took down the paper, but not before it was widely shared and fuelled a fair bit of vaccine hesitancy. While peer review at least it protects against easy-to-spot errors and misinterpretations.
The Cleveland Clinic study is such a preprint. But even if we put concerns about preprints aside, this one fundamentally does not support the things people are saying it does. It does not say that the bivalent booster increases the risk of catching COVID. The top line results show that the bivalent booster reduced COVID infections by 30 per cent. This result is supported by a recent that found the bivalent booster was better than the original vaccine and had an effectiveness of 59 per cent against hospitalization and 62 per cent against hospitalization or death.
The claim about vaccines making things worse stems from some of the secondary results and appears to fall victim to an epidemiological concept called The name comes from the convention that Table 1 in most papers presents the characteristics of patients and Table 2 looks at the relationships between various factors and the end result being studied.
Here鈥檚 an illustrative example. If you were studying whether people who carry lighters are at increased risk for lung cancer, you would have to adjust for smoking status. People who carry lighters are indeed more likely to get lung cancer. But if you adjust for smoking status, then obviously, it鈥檚 not whether someone carries a lighter per se that predicts lung cancer risk. Conversely, if you wanted to see whether smoking increased the risk of lung cancer (by comparing smokers to non-smokers), the presence or absence of a lighter in someone鈥檚 pocket shouldn鈥檛 affect the results all that much.
The Table 2 fallacy states that you should not assume that all factors should be adjusted for in the same way. If you want to study lighters, you should adjust for smoking. But if you want to study smoking, you don鈥檛 really need to adjust for lighter ownership. These relationships are not interchangeable.
So, if researchers want to study the benefits of the bivalent booster, they should adjust for prior vaccination. But the same formula cannot be used to tell you anything about the protective or harmful effect of prior doses that you adjusted for. It is a subtle but important mathematical principle that is often overlooked. Because if you want to study the number of prior doses, a much more important variable to consider is the timing since your last dose, which is probably more important.
